This work is of interest for clinical scientist, applied scientist and experts in musculoskeletal MR at 3.0T and 7.0T. MRI of subtle anatomical structures of the shoulder joints constitutes a challenge for diagnostic imaging due to spatial resolution constraints at 1.5 T and 3.0 T. Progress in ultrahigh field MR holds the promise to address this challenge with the ultimate goal to put the intrinsic sensitivity advantage at 7T into clinical use. Yet, some of the inherent advantages of UHF-MR are offset by practical obstacles including RF power deposition constraints. Recognizing this challenge this work examines relative RF power deposition for shoulder MRI with dual echo steady state imaging at 3.0T and 7.0T.

Volunteers (n=10, mean age 36.5 ± 8.51 years) were investigated at 3.0 T and 7.0 T using whole body MR scanners (3.0T, MagnetomVerio, 7.0T, Magnetom (Siemens, Erlangen Germany). At 3.0 T (1H 123Hz) a body coil was used for excitation while a 4 channel shoulder coil (Siemens, Erlangen, Germany) was used for reception. At 7.0 T (1H 297Hz) a 12 channel transceiver RF coil array dedicated for shoulder imaging was employed. The array comprises three modestly shaped sections to conform to an averaged shoulder. Each section contains 4 loop elements. The elements are organized in a 2 x 2 matrix. Electromagnetic (EM) field and SAR simulations were performed using Studio Suite 2012 (CST, Darmstadt, Germany) together with the voxel model Duke from the Virtual Family (ITIS Foundation, Zurich, Switzerland). A circular polarized mode with phases according to the angular position of the modules was employed. All elements were connected to multipurpose transmit/receive switch boxes with integrated low-noise preamplifiers. No specific tuning and matching or subject transmission field shaping was applied. For all volunteers dual echo steady state (DESS) was performed in normal operating mode [1] using the same parameters (TR=17.52ms. TE=5.41ms, slice thickness= 2.0 mm, acquisition time were TA=6:40min, FOV=200 x 200 mm2, matrix size=320 x 320, receiver bandwidth=155Hz/pixel, no fat saturation, slices per slab=40) at 3.0T and 7.0T. Local SAR limits were applied for the 12 channel transceiver array applied at 7.0 T. Global SAR limits were applied for the body coil transmission at 3.0 T. The flip angle was varied for each field strength ranging from 0.09 rad (5 degree) to the maximal flip angle supported by the SAR limits. Relative SAR was obtained from the SAR monitoring routine provided with the scanners operating software. Statistical evaluation was conducted with a paired Mann-Whitney-U-test. Using statistical package for the social sciences, (SPSS, version 23, Chicago, Illinois) and Microsoft excel.All volunteers were scanned at 3.0T and 7.0T using the same parameters for 3D DESS. At 3.0T a flip angle range of 0.09 rad to 1.571 rad (90 degree) was achieved before hitting the SAR limit. At 7.0T the maximum flip angle range was limited to 0.785 rad (45 degree) due to the RF power deposition limits. The flip angle comparison between 3.0T and 7.0T showed a SAR gain of approximately 2.6 for the local RF coil setup used at 7.0 T versus the body coil configuration employed at 3.0 T.Transferring musculoskeletal ultrahigh field MR from clinical research into clinical routine is at an early stage of the development process. Notwithstanding this challenge 7.0T MRI holds the potential to advance the capabilities of musculoskeletal imaging. To meet this goal, the suboptimal copy and paste approach to protocol migration from 3.0T to 7.0T needs to be supplanted application-targeted redesign of imaging protocols including RF power deposition considerations.