Introduction

Various RF-coil types and arrays have been introduced for cardiac magnetic resonance imaging (cMRI) in humans at UHF (e.g., microstrip line resonators^{1, 2}, Tx/Rx loop arrays^3–6, dipole antenna arrays^7–9, and combined dipoles and loop arrays^{10, 11}. However, pigs have quite different thorax shape than humans. The misuse of the human coils for performing cMRI in pigs will lead to a degradation of the SNR and $$$B_1^+$$$-field homogeneity due to less loading and suboptimal element dimensions. For cMRI, it is required to design a densely multi-row, multichannel array with good decoupling mechanism and geometrical loop configurations to enable RF-shimming^{12, 13} and parallel imaging along all standard and multi-oblique cardiac views. In this work, we report the design, simulation and testing of an 8-element anterior cardiac array combined with a rectilinear 8-element posterior array to form an 8Tx/16Rx dedicated pTx coil array to demonstrate the feasibility of cMRI in pigs at 7T.

Methods

The anterior array design was based on central 3-elements (1, 3, and 4) form together Y-shape loops, which were decoupled using a common conductor and three shared decoupling capacitors (SDC) (Fig. 1a). The surrounding 5-elements (2, 5, 6, 7, and 8) were chosen to be triangular in shape. The geometrical 8-loops distribution of anterior array was selected to be symmetric around the central Y-shape 3-elements to enable RF-shimming and parallel imaging along all standard and multi-oblique cardiac views. The anterior array was combined with a standard posterior array comprised of 8-loop elements based on a standard rectilinear design (Fig. 1b) to form an 8Tx/16Rx pTx dedicated pig array. The dimensions of the central Y-shape 3-elements were 4.3cm×7.53cm, 3.5cm×9.0cm, and 3.5cm×9.0cm, respectively. The dimensions of the elements were chosen to achieve a sufficient $$$B_1^+$$$-field penetration at about 10cm depth within the pig torso in order to create a balance between the individual 8-loop dimensions and the external dimensions of the array¹². The central Y-shape 3-elements (1, 3, and 4) were decoupled using two SDCs of ($$$C_1^d$$$ and $$$C_2^d$$$). The surrounding five elements 2, 5, 6, 7, and 8 were chosen to be triangular in shape. They are decoupled from the central Y-shape 3-elements using two SDCs of ($$$C_4^d$$$ and $$$C_5^d$$$) and gaps of 1.4cm and 1.2cm. The elements 5 and 7 and the identical mirrored elements 6 and 8 were decoupled using a SDC of ($$$C_3^d$$$). For each loop, two equal tuning capacitors ($$$C_1^t$$$, $$$C_2^t$$$, $$$C_3^t$$$, and $$$C_4^t$$$) were distributed equally in the gaps between the segment in order to have the element resonating at 297.2 MHz. The total external dimension for the anterior array was 19.1cm×23.1cm. The rectilinear posterior array was built using 4×2 rectangular symmetric elements configuration¹³. EM-simulations were carried out using CST-Microwave-Studio for array design and $$$B_1^+$$$-field optimization in a pig phantom ($$$\epsilon_{r}$$$= 59.0 and $$$\sigma$$$= 0.79S/m) (Fig. 1d). RF-circuit co-simulation was employed for good matching, tuning, and decoupling at 297.2 MHz¹⁴. An in-house developed dedicated pig body phantom (volume≈16L) was built to mimic approximately the thorax shape of the pig and to validate the $$$B_1^+$$$-field simulations (Fig. 1e). The dedicated pig array was tested in phantom and 46kg ex-vivo pig with the measurement of FA-maps, SNR-maps, and g-factor maps with high parallel imaging acceleration factors (R=2, 3 and 4). The acceleration was set in the left/right (L/R) direction. The coil array was tested in an ex-vivo experiment with one female 46kg pig (German Landrace, Heinrichs Tierzucht GmbH, Heinsberg, Germany). The animal was provided for an ex-vivo measurement, following its approved use in project 55.2 DMS 2532-2-664 (Regierung Unterfranken, Germany). All measurements were performed on a 7T whole-body 7T Siemens Magnetom Terra scanner (Siemens Healthineers, Erlangen, Germany) in pTx mode.

Results and Discussion

Fig. 2 illustrates the simulated central transversal $$$B_1^+$$$-field distribution and the measured FA-maps within the pig body phantom for the individual 8Tx-channels ($$$T_{x1}$$$-$$$T_{x8}$$$). The dedicated pig array generates spatial $$$B_1^+$$$-field distribution appearing for all 8Tx-channels. Good agreement between CST-simulations and MR-measurements was achieved for the 8Tx-channels. Fig. 3 demonstrates the simulated combined central transversal $$$B_1^+$$$-field and the measured FA-map, SNR-map, and g-factor maps within the pig phantom. There were some differences between simulated and measured FA-maps. Parallel imaging with an acceleration factor of R=4 was possible with the pig array to maintain the g-factor within the selected ROI at 1.18±0.12. Fig. 4 shows the noise correlation matrix, SNR-map, FA-map, and g-factor maps acquired using the pig array loaded with an ex-vivo 46kg pig. The measured SNR ± SD within the heart ROI of the pig was 36 ± 23. The dedicated pig array has good parallel imaging performance with R=4. The mean ± SD of the g-factor in the heart ROI was 1.14±0.07. High resolution ex-vivo long axis (LA) and short axis (SA) cardiac images were acquired with an in-plane resolution of up to 0.3mm×0.3mm using the dedicated pig array (Fig. 5).

Conclusion

In this work we demonstrated the feasibility of performing cMRI in pigs at 7T using an 8-elements array design based on central 3-elements Y-shape arrangement combined with a rectilinear 8-elements posterior array. The main achievement of the dedicated pig coil array was the possibility to obtain cardiac images with high in-plane resolution of up to 0.3mm×0.3mm.

Acknowledgements

This project is funded by the German Ministry of Education and Research (BMBF) with grant # 01EO1004 & 01EO1504.