Md Shahadat Hossain Akram1, Craig S. Levin2, Takayuki Obata1, Fumihiko Nishikido1, Eiji Yoshida1, and Taiga Yamaya1
1Molecular Imaging Center, National Institute of Radiological Sciences, Chiba, Japan, 2School of Medicine, Stanford University, Stanford, CA, United States
Synopsis
A prototype of a novel oval shape PET insert for
simultaneous body imaging with the MRI systems has been proposed in this study.
The smaller prototype has the minor axis and major axis of 14 cm and 21 cm. 16
copper shielded boxes are positioned on the periphery of the oval frame. The
shielding boxes were kept floating to let the RF field penetrate through the
gaps in between the 16 modules. To get the required RF field distribution
inside the oval PET, the gaps between the shielded modules were varied by doing
assumptions following the conformal electric phase angle methods. B1 maps and GRE and SE images were
taken and they have a good agreement with the results for without shielding
materials. We have found a reduced RF field value with increased noise in the
FOV which is mostly due to shielding materials. The images of GRE and SE have shape
distortions due eddy currents. Introduction
Dedicated PET inserts for the existing MRI systems have got a great deal of interest in recent years. But almost all of them are concentrated for human brain imaging or animal imaging. So far they can be categorized in three types of model: (a) cylindrical PET ring outside a cylindrical RF Tx/Rx coil [1]; (b) cylindrical PET ring integrated with Tx/Rx Birdcage coil [2]; (c) RF penetrable cylindrical PET ring – RF body coil for transmission and separate RF coil inside the PET ring for reception [3]. We are working for developing a whole body PET insert for the MRI systems: an oval shape RF penetrable PET insert would be a good solution to be inserted within the MRI bore for simultaneous imaging. It will make us possible to use the present concept of using body RF coil for transmission and surface coils for reception. A smaller prototype has been developed and experiment has been conducted considering the RF field (B1) distribution and, spin-echo (SE) and gradient echo (GRE) imaging.
Methods and Materials
In reference [3], eight shielded PET modules with similar dimensions were positioned in a cylindrical frame with equal gap in between the PET modules – the gaps are equidistant on the periphery of the cylinder. The PET modules were battery-powered to keep them floating from the RF ground that enables the transmission field of the body RF coil to enter through the gaps between the modules. Following cylindrical symmetry, a uniform RF field was created inside the floating PET ring.
To get the required RF transmission field inside the oval shape PET insert, the shield boxes dimensions and the gaps in between the shield boxes were different in this design. Our assumptions were followed by the conformal mapping of the equidistant electric phase angles [4] of circular cylindrical case to an elliptic case (Fig. 1(a)). Following the conventional PET design, to keep the equal number of detectors in each PET modules, in one design we kept the gap distribution equidistant on the periphery of an oval frame with unequal gaps in between the PET modules (Fig. 1(b)). Near the minor axis we kept smaller gap-width between the PET modules and near the major axis the gaps were larger - since conformal mapping shows dense gap-widths near the end of major axis than near minor axis.
The minor and major axes of the elliptic frame are 14 cm and 21 cm. We included 16 copper foil shielded PET modules. The complete model is shown in Fig. 2. Experiments were conducted in a 3 T clinical MRI system (Siemens Magnetom Verio) by using a homogeneous NaCl solution phantom. For the B1 field measurements, we implemented double angle method with the following sequence parameters: TR = 3s, TE = 12ms, Slice thickness (ST) = 5mm, 128 X 128 matrix, FOV =150mm (for transverse images) and 200mm (for coronal image). For the two images the excitation flip angles were 60o and 120o and, refocusing flip angle were 180o. The GRE imaging parameters: TR = 200ms, TE = 12ms, Average = 2, ST = 5mm, 128X128 matrix, FOV =150mm, Flip gangle = 60o. And the SE imaging parameters were: TR = 200ms, TE = 4ms, Average = 2, ST = 5mm, 128X128 matrix, FOV =150mm, Flip angle = 60o.
Results and Discussion
Fig. 2 summarizes the experimental results. In Fig. 2(a) we have given B
1 maps for center slice and 4 cm off-center slices. We compared the results with the B
1 maps for the case of without shielding boxes. Fig. 2(b) shows the coronal B
1 map with shield boxes. For comparison, the same color bar (shown with Fig. 2(b)) was used for all of these B
1 maps. The RF field due to shielding boxes was found to be in smaller field values compare to the without shield case. Fig. 2(c) shows comparative results of flip angle plot for the central line. For the shielded case, flip angle was lower due to the lower value of the RF field inside the shielded oval frame. This is due to the power loss related to the shielding currents and possible changes in the tuning and matching of the body RF coil. SNR is seen to be lower because of the noise generated in the shielding boxes, which is clearly seen in GRE and SE images in Fig. 2(c). For the shielding case both the images show shape distortions (in the lower part of the images) which could be mostly due to the eddy current effects generated in the shield boxes.
Acknowledgements
No acknowledgement found.References
[1] A. Kolb, et al, Eur Radiol 22 (2012) 1776–1788; [2] F. Nishikido, et al, Nucl Instru Method Phys Res A 756 (2014) 6–13; [3] P. Olcott, et al, Phys Med Biol 60 (2015) 3459–3478; [4] M. C. Leifer, Magn Reson Med, 38 (1997) 726-732.