4502

Retrospective transmission field (B1+) and sensitivity profile (B1-) correction for transceive surface RF coils: an empirical solution for RARE

Paula Ramos Delgado¹, Andre Kuehne², Ludger Starke¹, Jason M. Millward¹, Joao Periquito¹, Thoralf Niendorf^1,3, Sonia Waiczies¹, and Andreas Pohlmann¹

¹Berlin Ultrahigh Field Facility (B.U.F.F), Max Delbrück Center for Molecular Medicine, Berlin, Germany, ²MRI.tools GmbH, Berlin, Germany, ³Experimental and Clinical Research Center, a joint cooperation between the Charité Medical Faculty and the Max Delbrück Center for Molecular Medicine in the Helmholtz Association, Berlin, Germany

Synopsis

Improving the low signal-to-noise ratio (SNR) inherent to emerging MRI methods such as fluorine MRI is challenging. To enhance sensitivity, SNR-efficient pulse sequences such as RARE and cryogenically-cooled surface RF coils (CRP) are used. Transceive surface RF coils show variation in the excitation field (B₁⁺), impairing quantification. To compensate, previous studies have used an analytical signal intensity equation to perform a retrospective B₁⁺-correction. However, this is unfeasible for RARE due to the absence of such an equation. To overcome this challenge, we propose and validate a numerical method using experimental data acquired with a volume resonator (reference) and a ¹H-CRP.

Introduction

Fluorine (¹⁹F) MRI techniques support quantification but are limited by their inherently low signal-to-noise ratio¹ (SNR). To boost SNR, it is conceptually appealing to use SNR-efficient pulse sequences such as RARE^2-4 in combination with state-of-the-art cryogenically-cooled RF coils⁵ (CRP). However, ¹⁹F-CRPs are only available as transceive surface RF coils, which have a strong intrinsic spatial gradient in the excitation field (B₁⁺). As a result, both coil sensitivity and excitation flip angle (FA) are spatially dependent, severely hampering quantification. B₁⁺ mapping in combination with an analytical description of the signal intensity (SI) dependency on the FA permits a retrospective B₁⁺ correction⁶, compensating the effects of these FA variations on the SI. While this approach has been successfully applied to gradient echo techniques like FLASH, it is not feasible for RARE due to its more complex train of spin-echoes and stimulated echoes with error propagation⁷ for which no analytical SI equation exists. To address this shortcoming, we developed an alternative approach based on estimating the relationship between SI, FA and T₁ empirically. In this proof-of-concept study, data was acquired with a volume resonator (reference) and a ¹H-CRP to evaluate the feasibility of the proposed B₁ correction approach in test phantoms and ex vivo mouse brains. This method may be valuable not only for X-nuclei MR in which absolute SI is highly relevant, but also in conventional ¹H-MRI when accurate FAs are needed for well-defined T₁-contrasts and quantitative T₁-measurements.

Methods

Experiments were performed on a 9.4T animal MR scanner (Bruker BioSpin, Ettlingen, Germany).

B₁-mapping. A B₁⁺ map of the ¹H-CRP was computed on a uniform phantom using the double flip angle method^9-10. The approximation SI α B₁⁺/max(B₁⁺) · B₁^-/max(B₁^-) at low FAs was used to calculate the corresponding B₁^- map.

RARE signal intensity modelling. Aqueous solutions of gadolinium were used to prepare 12 NMR tubes with T₁ relaxation times ranging between 670 and 2850ms. RARE scans were acquired with a ¹H volume RF coil⁸ (TR=1000ms, ETL=8, BW=50kHz, centric encoding with flipback) and 12 reference powers to vary the FA between 35°-110°. The SI resulting from each FA-T₁-combination was obtained by averaging over ROIs in each NMR tube. A 6^th order polynomial was fitted to the experimental data to estimate SI=f(FA,T₁).

Image acquisition and T₁-mapping. The proposed approach was evaluated on a uniform phantom with two water-Gd-mixtures (15ml tube containing mixture-1 and two NMR tube inserts containing mixture-2) and a paraformaldehyde-fixed mouse head. RARE images were acquired (same parameters as above) using the volume resonator (reference) and the ¹H-CRP. T₁ maps were calculated using RARE (13 TRs from 100-18000ms).

Correction method. A B₁⁺-correction factor was calculated as the modelled RARE SI for perfect 90° excitation divided by the modelled RARE SI for the actual excitation FA. SI correction involved applying this B₁⁺-correction factor and dividing by the B₁^- map.

Evaluation. We compared the SI in the corrected CRP-image with that of the volume coil, along a line perpendicular to the RF coil surface.

Results

The RARE SI dependency on FA and T₁, as modelled with a polynomial fit to the experimental data, is illustrated as a surface plot in Fig.1. Good correspondence between the fit and the measured data points can be observed (RMSE=2.5170, R² = 0.9952). After B₁^- correction, the strong spatial SI gradient in the CRP-image was removed, yielding a rather homogeneous SI up to a depth of approximately 13.5mm, both for the uniform phantom (Fig.2) and the mouse brain (Fig.4). Plots of the normalized vertical SI profiles demonstrate the impressive increase in image homogeneity (Fig.3, Fig.5). While a meaningful SI quantification was unfeasible in the original image, the B₁-corrected RARE image permitted quantitation up to 13.5mm from the RF coil, sufficient to cover the entire mouse brain.

Conclusions

Here we developed a new method for B₁ correction when using RARE in combination with TX/RX-surface RF coils, such as state-of-the-art cryogenically-cooled probes. This proof-of-concept study demonstrates the feasibility of this B₁ correction approach, which is fundamentally limited by SNR constraints at larger distances from the RF coil. The proposed retrospective B₁ correction method for RARE may be valuable for conventional ¹H-MRI when accurate FAs are needed to achieve well-defined T₁-contrasts and quantitative T₁ measurements. This method is also highly relevant for quantitative MR of X-nuclei where absolute SI is limiting, as shown by the example of quantification of fluorine-labelled cells in neuroinflammation using a ¹⁹F-CRP.

Acknowledgements

This work was supported by the Deutsche Forschungsgemeinschaft to S.W. (DFG WA2804) and A.P. (DFG PO1869).

References

1. Waiczies, S, et al. Tracking methods for dendritic cells. ¹⁹F Magnetic resonance imaging, chapter 9, pp. 243-281.

2. Hennig, J, et al. RARE imaging: a fast imaging method for clinical MR. Magn Reson Med, 1986; 3: 823-833.

3. Mastropietro, A., et al. Optimization of rapid acquisition with relaxation enhancement (RARE) pulse sequence parameters for ¹⁹F-MRI studies. J. Magn. Reson. Imaging 2014; 40(1):162-170.

4. Faber, C., Schmidt, F. Chapter 1: Pulse sequence considerations and schemes. Fluorine Magnetic Resonance Imaging (2017).

5. Waiczies, S, et al. Enhanced fluorine-19 MRI sensitivity using a cryogenic radiofrequency probe: technical developments and ex vivo demonstration in a mouse model of neuroinflammation. Sci Rep 2017; 7: 9808.

6. Vernikouskaya, I, et al. In vivo ¹⁹F MRI quantification using B₁⁺/B₁^- correction. ISMRM 25th Annual Meeting & Exhibition, 2016.

7. Meara S J P, et al. Evolution of the longitudinal magnetization for pulse sequences using a fast spin-echo readout: application to fluid-attenuated inversion-recovery and double inversion-recovery sequences. Magn Reson Med, 2005; 54: 241-245.

8. Waiczies, H, et al. Visualizing brain inflammation with a shingled-leg radio-frequency head probe for ¹⁹F/¹H. Sci Rep., 2013; 3: 1280.

9. Akoka, S, et al. Radiofrequency map of an NMR coil by imaging. Magn Reson Imaging, 1993; 11:437-441.

10. Insko, EK, et al. Mapping of the radiofrequency field. J Magn Reson, 1993; A103:82-85.

Figures

Fig.1. RARE signal intensity (SI) model: (A) 3D-plot of the RARE signal intensity (SI) modelled as a function of the T₁ relaxation time and flip angle (6th order polynomial). (B) shows the SI vs FA plane and (C) the SI vs T₁ plane for selected FA and T₁ values. Each colored line depicts a different T₁ and FA, respectively. The dots show the measured data points. Fitted signal intensity data shows great resemblance with the originally measured data (RMSE = 2.5170, R² = 0.9952).

Fig.2. Uniform sample correction which consists of two different T₁ values for the NMR inserts (mixture-1) and 15ml tube (mixture-2), respectively. Column (A) shows the T₁ map of the sample and the B₁⁺ map of the ¹H-CRP. In column (B) the reference image acquired with the volume RF coil and the image to be corrected (¹H-CRP) expose the need for a B₁-corrected image. Quantification is severely hampered by the signal intensity gradient in the ¹H-CRP image. Column (C) shows the corrected ¹H-CRP image. To quantify the improvements, normalized profiles along the colored lines were plotted in Fig.3.

Fig.3. Normalized vertical profiles along the lines defined in Fig.2 show the improvements in signal intensity (SI) homogeneity achieved in the uniform phantom through a line across the 15ml tube (top) and through a line across the NMR tubes (bottom) with increasing distance from the ¹H-CRP surface. The quantification range was increased to approximately 13.5mm from the ¹H-CRP surface in the B₁-corrected image.

Fig.4. Ex vivo mouse brain correction. Column (A) shows the T₁-map of the mouse brain and the B₁⁺ map of the ¹H-CRP. In column (B) the reference image acquired with the volume RF coil and the image to be corrected (¹H-CRP) exposes the urgent need of a B₁-corrected image for quantification purposes in the brain region. Column (C) shows the corrected image. To quantify the improvements in signal intensity, vertical profiles along the colored lines were plotted in Fig.5.

Fig.5. Vertical profiles along the lines depicted in Fig.4 show (A) the improvements in signal intensity homogeneity achieved in the ex vivo mouse brain using the described B₁ correction method. (B) shows the comparison of the B₁-corrected image profile with the volume resonator. The similarity in the shapes of both curves can be observed.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)

4502