MRI was performed at 3 T (Prisma,
Siemens Healthcare) in three healthy subjects,
each at a different clinical site: University College London (UCL),
Max Planck Institute (MPI) for Biological Cybernetics in Germany,
and Tel Aviv University (TAU) in Israel. In addition,
one patient with one brain tumor patient with a high-grade glioma was scanned in Tuebingen.
All subjects provided written,
informed consent prior to participation and each study was approved by the local ethics committee.
Snapshot CEST imaging was performed with a 3D gradient echo readout optimized for 3 T ,
with 1.7x1.7x3 mm3 resolution and 220x178x54mm3 field-of-view.
Prior to each evaluation,
motion correction ,
B0 correction ,
and spectral denoising were applied.
Three different CEST contrasts were acquired and evaluated.
Amide proton transfer weighted (APTw) CEST
APT-weighted presaturation was achieved using a train of 20 Gauss-shaped pulses with B1= 2.2 μT,
pulse duration tp = 50ms,
and 55% duty cycle.
Images were acquired at 30 evenly spaced frequency offsets between +6.0 ppm and - 6.0 ppm,
with an unsaturated reference image acquired at 300 ppm.
Acquisition time was 4.1 minutes. After normalization of each CEST image by the reference image,
CEST contrast was evaluated using MTR asymmetry .
Nuclear Overhauser Effect (NOE) CEST
An optimized low-power presaturation to target spectrally selective amide- and nuclear Overhauser effect (NOE) CEST effects was achieved using a train of 80 Gauss-shaped pulses with B1= 0.6 μT,
pulse duration tp = 20ms,
and 50% duty cycle. Acquisition time was 6.5 minutes for 56 variably-spaced frequency offsets.
After normalization of each CEST image by the reference image,
Lorentzian fitting of the Z-spectrum was used to isolate background contributions and CEST contrasts .
Hydroxyl weighted (OH) CEST and dynamic glucose-enhanced (DGE) imaging
OH-weighted CEST presaturation was achieved using two Gauss-shaped pulses with B1= 3 μT,
pulse duration tp = 100ms,
and 100% duty cycle,
or by one adiabatic spin-lock pulse  with B1= 5 μT and 1.2 s pulse duration.
For a dynamic glucose-enhanced imaging baseline experiment,
no glucose was injected to measure stability of endogenous contrast over time. Acquisition time was 4.8 minutes for OHw-CEST with 38 evenly-spaced frequency offsets between +3.0 ppm and -3.0 ppm,
or 15.4 minutes for DGE-CEST with 25 repetitions of 5 interleaved frequency offsets (0 ppm,
and 1.5 ppm). After normalization of each CEST image by the reference image of each repetition,
the OH-weighted contrast was computed for each frequency offset,
and change from baseline (ΔDGE)  was computed in gray and white matter ROIs.