Be prepared, stay tuned and keep’em separated: CEST MRI in the human brain at 9.4 Tesla

Purpose/Introduction: Chemical exchange saturation transfer (CEST) MRI benefits from high magnetic field strengths due to the increased spectral resolution [1]. Thus CEST at 9.4 T is tempting, however, presaturation at 9.4 T faces SAR limits as well as B1 inhomogeneities, moreover the selective saturation faces the stronger B0 inhomogeneity at 9.4 T. Thus we aimed for a CEST approach at 9.4 T with good CEST preparation, field inhomogeneity correction, and access to several well-separated CEST contrasts. Subjects and Methods: In vivo CEST-imaging was performed on a 9.4 T whole body MRI tomograph (Siemens) using a 2-D GRE sequence on a healthy volunteer with informed consent and approval by the local ethics committee. A custom-built head coil [2] was used for signal transmission/reception (16 transmit/31 receive channels). GRE CEST images at 58 unevenly distributed offsets, were obtained after saturation by a train of 150 Gaussian-shaped RF pulses (tpulse = 15 ms, tdelay = 10 ms, tsat = 3.75 s) at B1 = 0.5 lT and 0.8 lT [3]. Four CEST contrasts and the ssMT were calculated by fitting the Z-spectra with a 6-pool Lorentzian model [3]. Using an AFI B1 mapping [4], contrasts were corrected for B1-inhomogeneities [5]. Using the minimum of the Z-spectrum as a B0 deviation map, Z-spectra were pixel-wise B0 corrected. Results: Residual B0 deviations ( ±0.3 ppm = 120 Hz) were correctable due to high offset sampling; also B1 inhomogeneities (down to 50 ) can be flattened with the previously proposed Z-B1-correction algorithm [4] (Fig 2b–e). Both the increased SNR (Fig. 1c) but also the longer T1 times at 9.4 Tesla (Fig 1d) help to increase the CEST sensitivity. This and the increased spectral resolution allows to separate the different CEST peaks; APT shows the typical increased signals in GM, NOE shows the typical correlation with ssMT. In addition, the amine peak at 2.8 ppm, hardly visible at 7T [3, 5, 6], could here be detected and evaluated to generate a CEST@2.8 ppm map (Fig. 2c). Thus within 15 min of scan time 4 different CEST contrast could be obtained within SAR limits that can be used to detect metabolic and physiologic changes in various pathologies. Discussion/Conclusion: We presented GRE based CEST approach at 9.4 T that allows for CEST-prepared MRI with decent SNR ([200) and within SAR limits. Incorporated B0 and B1 correction allows for reconstruction of well resolved Z-spectra and generation of different homogeneous CEST contrast maps in the human brain separately