3912

Quality Assurance Phantoms and Procedures for UHF MRI ‒ The German Ultrahigh Field Imaging (GUFI) Approach
Maximilian N. Voelker1, Oliver Kraff1, Eberhard Pracht2, Astrid Wollrab3, Andreas K. Bitz4, Tony Stöcker2, Harald H. Quick1,5, Oliver Speck3,6, and Mark E. Ladd1,4

1Erwin L. Hahn Institute for Magnetic Resonance Imaging, University of Duisburg-Essen, Essen, Germany, 2German Center for Neurodegenerative Diseases (DZNE), Bonn, Germany, 3Otto-von-Guericke-University, Magdeburg, Germany, 4Medical Physics in Radiology, German Cancer Research Center (dkfz), Heidelberg, Germany, 5High Field and Hybrid MR Imaging, University Hospital Essen, University of Duisburg-Essen, 6Leibniz Institute for Neurobiology, Magddeburg, Germany

Synopsis

The German Ultrahigh Field Imaging network (GUFI, www.mr-gufi.de) is a user group of 13 German and neighboring sites that all operate a UHF (7T or 9.4T) MRI system. Due to the lack of common quality assurance (QA) procedures for UHF, GUFI started an initiative to unify QA procedures at these sites. A QA phantom and measurement protocol were developed especially for UHF that is currently being rolled out to all member sites. The QA data allow monitoring of individual system performance based on long-term data analysis or by comparison to pooled data from all sites.

Purpose

In this study we developed a quality assurance (QA) phantom and measurement procedure for UHF MRI. We collected QA data from four sites equipped with 7T systems and compared their key system-related performance parameters. The proposed method should allow assessment of system performance and comparison to other UHF MRI scanners.

Material and Methods

The QA phantom needs a stable filling material with dielectric properties similar to human tissue. Therefore, a dilution series for polyvinylpyrrolidone (PVP) and salt in water was mixed to find a recipe for the modification of the dielectric properties suitable for both 7T and 9.4T MR systems. This recipe was used to calculate the concentrations needed for the phantom fluids. The housing of the phantom was made with two compartments. The head-and-neck part is formed from a PMMA tube terminated on the two ends with a half shell and a plate, respectively (Fig 1). This container was fitted into a widely available 1TX/32RX-channel RF head coil (Nova Medical, MA) and filled with a gel of water, PVP, agar, and salt to emulate brain tissue. The second part of the phantom was designed to emulate the load of the human shoulders on the RF coil. Therefore, a rectangular container was fitted adjacent to the coil and head-and-neck part (Fig 1). This container was filled with a solution of PVP and salt to emulate muscle tissue. Four of the QA phantoms were fabricated, and QA analysis has started at four different UHF sites. All sites are equipped with a whole-body 7T MRI (Siemens Healthcare), but some major hardware components, i.e. the magnet, the gradient coil, and the RF coil differ between the sites. (Fig 2 / Table 1). A measurement protocol was created to check for RF coil performance as well as gradient and system stability; in addition, the protocol was designed to verify the stability of the phantom filling material. RF coil performance was measured with B1, SNR, and coupling measurements like noise correlation or SNR maximum position analysis (Fig 3). Using noise measurements, the RF system was checked for unwanted noise sources and the gradients / system were checked for RF spikes. The system was stressed with high-duty-cycle EPI bold imaging to capture stability parameters, e.g. signal drift or fluctuation according to the recommendations of the fBIRN consortium1. Relaxometry with multi-echo spin-echo and inversion recovery spin-echo measurements were appended to confirm the long-term stability of the phantom. From the acquired images, parameters such as central SNR or flip angle, noise correlation, system drift, and fluctuation or relaxivity were calculated; the data were analyzed across longitudinal measurements and between sites. Before delivery, all phantoms underwent a complete QA measurement with this QA protocol at the same site (Fig 4).

Results

Based on dielectric and relaxation properties, the phantom fluid was suitable to emulate the desired tissues. The phantoms had similar relaxation behavior at all sites (mean: T1 = 680ms, all deviations <= 10%, T2 = 58ms, dev. <= 3%). The mean relative permittivity of all phantoms was 55.5 (dev. <= 4%), and the mean conductivity was 0.67 S/m (dev. <= 8%). Analysis across sites revealed high agreement regarding basic image parameters such as SNR and achievable flip angles if an RF coil of the same type was used. Differences caused by the different models of the gradient and RF coils can easily be identified (Fig 4). The inter-site differences are in the same range as the differences found in longitudinal data analysis (Fig 5). System stability parameters revealed differences between the sites, and longitudinal data confirmed these findings (Fig 5).

Discussion and Conclusion

At UHF the dielectric properties of the measurement objects play an important role due to artifacts arising from shorter wavelengths. Therefore the QA phantom had to match the dielectric properties of the emulated tissue as well as the relaxation properties. The phantom proved suitable for QA as these parameters create with the simple geometry an adequate coil load in a stable and storable form. A comprehensive measurement protocol was created to check system performance. By pooling the QA data of four UHF sites additional benefit and analysis power was gained. With the successful implementation of this QA procedure further sites will be included. All 13 sites of the GUFI network (www.mr-gufi.de) with ten 7T and two 9.4T MRI systems will soon be equipped with phantoms to benefit from the proposed QA method.

Acknowledgements

The research leading to these results has received funding from the German Research Foundation (DFG) / project German Ultrahigh Field Imaging / Grant n. LA 1325/5-1

References

1. Friedman et al., JMRI 23:827–839 (2006)

Figures

Fig.1: The GUFI QA phantom consists of two PMMA containers. The head-and-neck part is filled with a gel of water (62%), PVP (36%), agar (1.5%), and salt (1.0%). The shoulder part is filled with a solution of PVP (31%) and salt (0.9%). The housing provides a stable and simple geometry. The filling provides realistic dielectric properties combined with a realistic T2 relaxation time (57ms) and acceptable T1 relaxation time (650ms). Desirable features such as gel-like consistency, a single peak in the proton signal spectrum, and a stable form without the need for poisonous additives favor the use of these ingredients.

Fig. 2 / Table 1: Hardware configuration at the different sites.

Fig. 3: Excerpt of data analysis. In A) - D) single-channel SNR maps were used to calculate the channel-wise SNR maximum and mean. The position of the SNR maximum was used as an additional measure for coil decoupling state and performance. E) shows a transversal SNR map gained from the combined image. The noise correlation F) was used to asses coil performance by calculating differences to previous measurements. The RF noise spectrum G) gives information whether unwanted RF sources are present. System stability was tested with high-duty-cycle EPI measurements H), and stability parameters such as signal drift were calculated.

Fig. 4: The QA parameters of all phantoms were compared with reference measurements taken at Site 1 before shipping to other sites. Certain major imaging components differ between the 7T systems at the sites, but all systems are from the same vendor. The different gradient coils cause distortions due to linearity differences as seen in the non-distortion-corrected coronal actual flip angle images. The 24RX-channel version of the RF coil used at Site 4 had lower transmit efficiency, especially in the inferior phantom region. 320V reference amplitude was used at Site 4 compared to 260V for all 32RX-channel-coil sites.

Fig. 5: Excerpt of the longitudinal data analysis. For fBIRN analysis the second and the third EPI time series (250 frames each) at all sites are presented to allow comparison of the systems in a defined warm-up state. The third series was acquired with half of the nominal flip angle and has lower SNR and signal drift but higher radius of decorrelation values. Site 1 had a system malfunction during the measurements on 2016/10/28 (marked with red arrows) with unusually low SNR and high signal fluctuations (probably caused by loose coil contacts).

Proc. Intl. Soc. Mag. Reson. Med. 25 (2017)
3912