An 8Tx/32Rx RF Coil for 7T UHF Body MRI
Stefan HG Rietsch1,2, Stephan Orzada1, and Harald H Quick1,2

1Erwin L. Hahn Institute for MR Imaging, University of Duisburg-Essen, Essen, Germany, 2High Field and Hybrid MR Imaging, University Hospital Essen, Essen, Germany

Synopsis

In order to allow for improved SNR and higher acceleration during image acquisition in the body at 7T, we present a coil with 8Tx/Rx microstrip line elements with meanders and 24Rx loop elements. This coil comprises 8 building blocks each consisting of one Tx/Rx element and 3 overlapping loops which are actively detuned during transmit. With about -19 dB reflection and an average decoupling of more than -30 dB, the SNR can be boosted by about 21% in the abdomen. Evaluation of g-factors as well as in vivo images of healthy volunteers in both abdomen and heart show promising results.

Purpose

Imaging in the human torso (thorax, abdomen, pelvis) at 7T MRI is challenging due to inhomogeneous radio frequency (RF) excitation and limited signal penetration depth. Using multiple transmit channels and meander elements (ME)1 are one way to tackle these problems. An expansion to a higher number of receive coil elements would potentially allow for higher SNR and acceleration factors2. In this work we present a RF coil with 8 transmit/receive (Tx/Rx) channels and 24 receive-only (Rx) channels (8Tx/32Rx) tailored for body imaging at 7T UHF MRI.

Material and Methods

Each of the eight building blocks (Figure 1A) of the 8Tx/32Rx coil consists of one Tx/Rx ME and three Rx loops oriented along the z-direction. The loops are overlapping and active detuning is provided by PIN diodes and a serial inductance in parallel to the capacitor close to the feeding point. Four of the building blocks are positioned in the basis below the patient (Figure 1B) and another four are tightly fitted on top of the patient with a belt (Figure 1C). All housings are made of polycarbonate and rounded edges guarantee optimized patient comfort. Custom made preamplifier boards enable preamp decoupling for the Rx loops and the connection to the system is accomplished by Tim cables and connectors (Total Imaging Matrix, Siemens Healthcare GmbH, Erlangen, Germany) for easy plug and play on the 7T whole-body research MR system (Magnetom 7T, Siemens Healthcare GmbH). The ME elements are connected to a custom built 8 channel Tx/Rx switchbox.

Results and Discussion

With Q-ratio values of 2.67 ± 0.10 the Rx loops work in the body noise dominated regime. Figure 2A depicts the S-parameter matrix measured for the torso on a male volunteer (179 cm, 70 kg). Due to the fixed overlap between neighboring loops, the maximum next neighbor coupling is S12 = -8.7 dB. On average, the coil performs well concerning reflection (ME: Sxx = -18.1 ± 1.4 dB; loops: Sxx = -19.4 ± 7.1 dB) and inter element coupling (ME: Sxy = -37.0 ± 8.5 dB; loops: Sxy = -44.1 ± 11.6 dB). The noise correlation matrix is given in Figure 2B. The preamplifier decoupling was -9.0 ± 2.4 dB for the ME and -9.0 ± 2.3 dB for the loops. By dividing SNR maps obtained with all elements receiving the signal by SNR maps using only the 8 Tx/Rx ME, maps of SNR gain quantify the boost provided by the loops (Figure 3). Evaluation was critical in the center of the slice because the sequence used a threshold for low intensity areas. Consequently, it was difficult to achieve reasonable SNR map values there. If only the areas are included where a SNR gain is measurable the average gain is about 21%. Quantification of the additional acceleration capabilities that are offered by the 24 loops are demonstrated in tabular form by a g-factor comparison (Figure 4) between image acquisition using only the 8 Tx/Rx ME and image acquisition using the loops in addition. For example, an acceleration in anterior-posterior (A-P) direction of R = 2 implies a maximum g-factor of max(g) = 2.31 when only the ME are used. Adding the loops allows for an acceleration of R = 3 in A-P and an additional acceleration of R = 2 in head-feet (H-F) direction with a maximum g-factor of max(g) = 2.21. Consequently, the maximum g-factors max(g) confirm a paramount boost when the loops are included which can be used to shorten acquisition time or to increase spatial and/or temporal resolution. In vivo measurements in the pelvis of a healthy volunteer (male, 33y, 176 cm, 80 kg) were acquired in transversal orientation using a 2D HASTE sequence with TIAMO3 (1.56x1.56x5 mm3, TR/TE = 1500/68 ms, nominal FA = 180°) for acceleration factors of R = 2 (Figure 5A), R = 4 (Figure 5B) and R = 6 (Figure 5C). As the images show, R = 4 is still applicable without hampering image quality by folding artifacts. In vivo measurements of the human heart during breath hold (healthy male volunteer, 39y, 186 cm, 80 kg) using a 2D FLASH sequence (1.0x1.0x3 mm3, R = 4, TR/TE = 46.6/5.2 ms) show good image quality and homogeneous intensity distribution (Figure 5D,E).

Conclusion

The presented 8Tx/32Rx RF coil offers promising first in vivo images with increased SNR as well as encouraging acceleration capabilities. Further evaluation will include image protocol optimization for high resolution UHF MRI in different body regions (thorax, abdomen, pelvis) and a comparison to the 8ch Tx/Rx RF body coil1.

Acknowledgements

No acknowledgement found.

References

1. Orzada S, QuickHH, Ladd ME, Bahr A, Bolz T, Yazdanbakhsh P, Solbach K, Bitz AK. A flexible 8-channel transmit/receive body coil for 7 T human imaging. Proc. Intl. Soc. MRM 15, #2999 (2009)

2. Roemer PB, Edelstein WA, Hayes CE, Souza SP, Mueller OM. The NMR Phased Array. MRM 16, 192-225 (1990)

3. Orzada S, Maderwald S, Poser BA, Bitz AK, QuickHH, Ladd ME. RF Excitation Using Time Interleaved Acquisition of Modes (TIAMO) to Address B1 Inhomogeneity in High-Field MRI. MRM 64:327–333 (2010)

Figures

Figure 1: The 8Tx32Rx coil consists of 8 building blocks (Figure 1A) which each comprise one Tx/Rx meander element and 3 overlapping Rx loops. In the bottom part (B) four of those building blocks are implemented. The upper four building blocks are integrated in boxes (C) that are flexibly connected.

Figure 2: The S-parameter matrix (A) for the meander elements (#1-8) and loops (#9-20 top, #21-32 bottom) indicates a next neighbor coupling between the loops of S12 = -12.0±2.3 dB while the maximum coupling between loops and ME is -20.5 dB. The noise correlation matrix is shown in B.

Figure 3: SNR gain maps in transversal orientation obtained by division of SNR maps acquired with both the 8Tx/Rx ME and the 24Rx loops, and SNR maps acquired with the 8Tx/Rx ME alone. As soon as the color differs from the background the 24Rx loops boost the SNR.

Figure 4: Maximum g-factors max(g) for different acceleration factors R. First, only the 8Tx/Rx meander elements were used. Adding the 24 loops (8Tx/Rx meander elements + 24 loops) provides distinct acceleration capabilities. This could be used to dramatically shorten the acquisition time or to increase spatial and/or temporal resolution.

Figure 5: In vivo imaging of the pelvis (2D-HASTE-TIAMO, 1.56x1.56x5 mm³, TR/TE = 1500/68 ms, FAnominal = 180°) with R = 2 (A), R = 4 (B) and R = 6 (C) and the human heart during breath hold (2D-FLASH, 1.0x1.0x3.0mm³, R = 4, TR/TE = 46.6/5.2ms, TA = 20s).



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
2131