4429

Signal interferences in Actual Flip Angle Imaging (AFI) of polyvinylpyrrolidone (PVP) solutions

Niklas Himburg¹, Max Lutz¹, Lorenz Mitschang¹, Jan Gregor Frintz¹, and Sebastian Schmitter^1,2,3
¹Physikalisch-Technische Bundesanstalt, Braunschweig and Berlin, Germany, ²Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany, ³University of Minnesota, Center for Magnetic Resonance Research, Minneapolis, MN, United States

Synopsis

Keywords: Quantitative Imaging, High-Field MRI, B1+ mapping, Actual Flip Angle Imaging (AFI)

Motivation: Polyvinylpyrrolidone (PVP) solutions have dielectric properties similar to human tissue and therefore are widely used for validating electromagnetic simulations in phantoms. For validating the flip angle in such applications, actual flip angle imaging (AFI) is highly suitable.

Goal(s): This work investigates measurement errors of AFI with PVP solutions.

Approach: Different PVP/H2O ratios are investigated at 3T, with different RF-spoiling phase increments, and with NMR spectroscopy.

Results: NMR spectra show the existence of off-resonant PVP signals. Integrating two isochromat ensembles with different frequencies into an EPG-simulation qualitatively explain the measured signals and resulting flip angle errors in AFI.

Impact: Understanding the cause of unexpected measurement errors of actual flip angle imaging with polyvinylpyrrolidone solutions will improve their use for the validation of new B₁⁺ mapping techniques or new transmit arrays.

Introduction

MRI at ultra-high fields (≥7T) yields heterogeneous B₁⁺ fields leading to position-dependent flip angle (FA) distributions¹. A fast and frequently used 3D-sequence for measuring FAs is actual flip-angle imaging² (AFI). In phantoms, AFI is used for accurate quantifications e.g., to validate B₁⁺ fields of UHF transmit arrays^3,4. Suitable phantom fillings with tissue-equivalent dielectric properties (ε, σ) are polyvinylpyrrolidone (PVP) solutions⁵ with PVP/water fractions of 30-50% as in^4,6. Own AFI measurements with certain RF spoil-phase increments (Φ₀), however, showed considerable deviations from the expected FAs without the presence of ghosting artefacts as reported in ref⁷.
In this work, we investigate the accuracy and systematic errors of AFI in PVP phantoms. Therefore, we extended the approach from Yarnykh⁸ to cover multiple precession frequencies. This is done using extended phase graph⁹ (EPG) simulations and phantom validation measurements.

Methods

AFI acquisition: All scans were performed at 3T (Verio, Siemens) using a 32-channel head coil. A non-selective 3D Cartesian AFI was applied with: field-of-view(FOV)=256x128x160mm; resolution=2x2x5mm; TE/TR1/TR2=1.90/25/125ms; nominal FA=60°; spoiler-gradient-areas (read direction): A_G1/A_G2=117.5/587.5mT$$$\cdot$$$ms/m. The phantom consists of four pairs of tubes: four water-filled tubes (termed W1-W4) with Gadolinium-based contrast agent (ADC=1.9$$$\,\cdot$$$10^-3mm²/s at 293K) and four PVP/water-filled tubes (PVP1=23% weight (wt) PVP, ADC=0.99$$$\,\cdot$$$10^-3mm²/s; PVP2=38wt.%PVP, ADC=0.53$$$\,\cdot$$$10^-3mm²/s; PVP3=44wt.%PVP, ADC=0.35$$$\,\cdot$$$10^-3mm²/s; PVP1=50wt.%PVP, ADC=0.24$$$\,\cdot$$$10^-3mm²/s at 293K). Pairs have similar T1-values (W1/PVP1=1230/1210ms; W2/PVP2=760/680ms; W3/PVP3=550/500ms; W4/PVP4=420/380ms).

AFI measurements were performed with Φ₀ ranging from 0°-160° in 20°-steps. FAs inside each tube were averaged.

FA Reference: FA reference maps were acquired using the same setup and an accurate but slow preparation-pulse method¹⁰ (PEX) (FOV=256x128x5mm; resolution=2x2x5mm; TE/TR=3.40/8000ms). RF voltages of the rectangular preparation pulses were changed in steps of 50V (range 0-350V). FAs were averaged in each tube.

NMR Spectroscopy: ¹H-NMR spectra of two PVP/D2O samples (20, 50wt.%PVP, T=298K) were acquired at a 1.88T spectrometer (Fourier80, Bruker). T1-times of the PVP resonances were measured by inversion recovery (TI=0-200ms, 20 timesteps).

EPG Simulation: Two 1D-EPG simulations, including diffusion, were performed:
i) using parameters (T1/T2/ADC) for PVP1-PVP4 and a single precession frequency to simulate the AFI signals S_1/2 for different Φ₀. FAs of the simulated RF-pulses were set to the FAs of the reference method.

ii) For PVP1 showing the strongest effect, simulations for two precession frequencies with difference ∆ν were performed and signals were combined by:
$$S_{1/2}=S_{1/2,a}(T1_a,T2_a,ADC)+w\,\cdot\,S_{1/2,b}(T1_b,T2_b,ADC,\Delta\nu)e^{i(\Delta\nu\,\cdot\,TE+\varphi)}$$ with a weighting factor w and an additional phase term (Δν$$$\cdot$$$TE plus optional fitting phase φ) for S_1/2,b. EPG simulations of S_1/2,awere performed with T1/T2/ADC as in i). For S_1/2,b (Δν=320Hz), values of T1_b/T2_b=100/6ms were used to resemble the PVP resonance from the 20wt.% PVP ¹H-NMR spectrum.

Results and Discussion

Fig.1 shows an AFI FA map with Φ₀=40°. Deviations from the nominal FA (60°) are small (0.5-9.0%) for all tubes except PVP1(39%) and PVP2(18%).
Fig.2 displays the effect of different spoil increments on AFI (FA(Φ₀)) for the four tube pairs. Qualitatively, curves for water (W1-4) and PVP4 appear symmetric around Φ₀=90°. All other PVP curves are asymmetric with a minimum at Φ₀=60° and decreasing minimum values for decreasing PVP content.
Fig.3 provides simulated and measured spoiling curves of PVP1/W1 together with the preparation-based reference scan. The single precession frequency approach was used as in ref^8,11. Simulated and measured curves match qualitatively for W1 and agree with published results for water^8,11. However, EPG simulations for one precession frequency do not reflect the asymmetry of PVP1-3. Therefore, their deviations from the nominal FA in AFI maps cannot be explained by spoiling-related errors alone.
¹H-NMR spectra of the PVP/D2O solution show two PVP resonances besides the residual water peak at 4.4ppm. PVP peaks of the 50wt.% PVP/D2O sample are broadened (FWHM/T2*=2.5ppm/1.6ms) compared to the 20wt.% sample (FWHM/T2*=0.7ppm/5.7ms). Measured T1-values are T1_20%=100ms and T1_50%=170ms. Thus, the short T1 in combination with high FAs in AFI are expected to impact the resulting signals.
Fig.5 displays the measured PVP1 curve and the EPG simulation considering two precession frequencies (one for water and one combined PVP frequency) and relative weights. By adjusting w and fitting φ≠0, the simulated curve shows good agreement with the measured one. Therefore, combining the AFI signals of multiple EPG simulations with different precession frequencies, enables the reproduction of the measured PVP spoiling curves. Due to the small T2*-times of the PVP resonances, this interference effect will vanish with increasing TE. Furthermore, higher gradient spoiling moments reduce this effect.

Conclusion

Under certain RF spoiling conditions, signals resulting from the PVP peaks must be considered for AFI-based FA validation measurements in PVP solutions. Taking this into account by increasing TE should allow for reliable AFI mapping in arbitrary PVP solutions even when using usual recommendations⁸ for choosing Φ₀.

Acknowledgements

The authors thank Małgorzata Marjańska for helpful discussion.

References

1. Ladd ME, Bachert P, Meyerspeer M, Moser E, Nagel AM, Norris DG, Schmitter S, Speck O, Straub S, Zaiss M. Pros and cons of ultra-high-field MRI/MRS for human application. 2018;109:1-50.

2. 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(1):192- 200.

3. Woo MK, DelaBarre L, Waks M, Radder J, Choi U-S, Lagore R, Ugurbil K, Adriany G. A 16-Channel Dipole Antenna Array for Human Head Magnetic Resonance Imaging at 10.5 Tesla. Sensors. 2021; 21(21):7250.

4. Schmidt S, Ertürk MA, He X, Haluptzok T, Eryaman Y, Metzger GJ. Improved ¹ H body imaging at 10.5 T: Validation and VOP-enabled imaging in vivo with a 16-channel transceiver dipole array. Magn Reson Med. 2023 Sep 13. doi: 10.1002/mrm.29866. Epub ahead of print.

5. Ianniello C, de Zwart JA, Duan Q, Deniz CM, Alon L, Lee JS, Lattanzi R, Brown R. Synthesized tissue-equivalent dielectric phantoms using salt and polyvinylpyrrolidone solutions. Magn Reson Med. 2018;80(1):413-419.

6. Voelker MN, Kraff O, Pracht E, Wollrab A, Bitz AK, Stöcker T, Quick HH, Speck O, Ladd ME. Quality Assurance Phantoms and Procedures for UHF MRI ‒ The German Ultrahigh Field Imaging (GUFI) Approach. Proceedings of the 25th Annual Meeting of ISMRM, Abstract 3912.

7. Samsonov A, Velikina J, Yarnikh V. Accurate B1 Mapping with Actual Flip Angle Imaging (AFI) in the Presence of Fat. Proceedings of the 31st Annual Meeting of ISMRM, Abstract 0668.

8. Yarnykh VL. Optimal radiofrequency and gradient spoiling for improved accuracy of T1 and B1 measurements using fast steady-state techniques. Magn Reson Med. 2010;63(6):1610–1626.

9. Weigel M. Extended phase graphs: dephasing, RF pulses, and echoes - pure and simple. J Magn Reson Imaging. 2015 Feb;41(2):266-95.

10. Seifert F, Wübbeler G, Junge S, Ittermann B, Rinneberg H. Patient safety concept for multichannel transmit coils. J Magn Reson Imaging. 2007 Nov;26(5):1315-21.

11. Nehrke K. On the steady-state properties of actual flip angle imaging (AFI). Magn Reson Med. 2009 Jan;61(1):84-92.

Figures

Figure 1: Transversal FA map of eight tubes filled with water (W1-4) or a water/PVP mixture (PVP1-4, PVP ratio: 23-50wt.%) acquired with the AFI sequence (Φ₀=40°). FAs outside the tubes are masked. The mean reference FAs in the tubes are: PVP1=58.3°, PVP2=60.1°, PVP3=57.9°, PVP4=57.4°, W1=63.4°, W2=61.2°, W3=60.3°, W4=58.6°. Mean deviations of the FA inside the tubes from the nominal FA=60° are maximal for PVP1 (39%).

Figure 2: RF Spoil increment dependency of the AFI measurements of tubes a) W1 and PVP1(23wt.% PVP ratio, ε=66.4, σ=0.62S/m), b) W2 and PVP2(38wt.% PVP ratio, ε=57.8, σ=0.60S/m), c) W3 and PVP3(44wt.% PVP ratio, ε=53.3, σ=0.55S/m), d) W4 and PVP4(50wt.% PVP ratio, ε=50.1, σ=0.56S/m). The dotted lines indicate FA reference values acquired with the preparation-pulse method (PEX). The W1-4 spoiling curves are symmetric around Φ₀=90°. Qualitatively, the PVP curves appear increasingly asymmetric with decreasing PVP ratio.

Figure 3: EPG simulations of a) PVP1 and b) W1 with only one considered precession frequency over different spoil increments. The simulation results for the W1 tube quantitatively match the measured FAs. The simulated spoiling curve for PVP1 does not show the asymmetry of the measured data.

Figure 4: ¹H-NMR spectra of two PVP/D2O samples with 20wt.% and 50wt.% PVP ratio. The peaks around 4.4ppm originate from residual water in the sample. Two additional resonances from the PVP are visible at 3.0ppm and 1.6ppm in both spectra. The PVP resonances in the 50wt.% sample are broadened compared to the 20wt.% sample with a Full-Width at Half Maximum (FWHM) of FWHM₅₀$$$\approx$$$2.5ppm (T2*=1.6ms) and FWHM₂₀$$$\approx$$$0.7ppm (T2*=5.7ms).

Figure 5: Double precession frequency EPG simulations of the PVP1 spoiling curve a) without additional phase φ≠0 and weighting factor according to the Proton number ratio (N_PVP/N_H2O), b) with φ and w adjusted. Taking only one additional precession frequency into account already leads (with φ, w adjusted) to good agreement of the simulated and measured spoiling curve of PVP1.

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

4429

DOI: https://doi.org/10.58530/2024/4429