Temporal B0 field variation effects on MRSI of the human prostate at 7 T and feasibility of correction using an internal field probe.

Spectral degradations as a result of temporal field variations are observed in MRSI of the human prostate. Moving organs generate substantial temporal and spatial field fluctuations as a result of susceptibility mismatch with the surrounding tissue (i.e. periodic breathing, cardiac motion or random bowel motion). Nine patients with prostate cancer were scanned with an endorectal coil (ERC) on a 7-T MR scanner. Temporal B0 field variations were observed with fast dynamic B0 mapping in these patients. Simulations of dynamic B0 corrections were performed using zero- to second-order shim terms. In addition, the temporal B0 variations were applied to simulated MR spectra causing, on average, 15% underestimation of the choline/citrate ratio. Linewidth distortions and frequency shifts (up to 30 and 8 Hz, respectively) were observed. To demonstrate the concept of observing local field fluctuations in real time during MRSI data acquisition, a field probe (FP) tuned and matched for the (19) F frequency was incorporated into the housing of the ERC. The data acquired with the FP were compared with the B0 field map data and used to correct the MRSI datasets retrospectively. The dynamic B0 mapping data showed variations of up to 30 Hz (0.1 ppm) over 72 s at 7 T. The simulated zero-order corrections, calculated as the root mean square, reduced the standard deviation (SD) of the dynamic variations by an average of 41%. When using second-order corrections, the reduction in the SD was, on average, 56%. The FP data showed the same variation range as the dynamic B0 data and the variation patterns corresponded. After retrospective correction, the MRSI data showed artifact reduction and improved spectral resolution. B0 variations can degrade the MRSI substantially. The simple incorporation of an FP into an ERC can improve prostate cancer MRSI without prior knowledge of the origin of the dynamic field distortions.

Composite slice-selective adiabatic excitation for prostate MRSI.

Higher magnetic field strengths, such as 7 T, offer increased spectral resolution and higher signal-to-noise ratio. These properties can be very advantageous for MRSI. In particular, signals that generally overlap at lower fields, such as choline, polyamines and creatine, can be resolved at 7 T. However, higher magnetic field strengths suffer from strong radiofrequency (RF) field nonuniformities. These nonuniformities become even stronger when using surface transceivers, such as an endorectal coil for prostate imaging. In order to obtain uniform excitations for accurate MRSI measurements, adiabatic sequences are therefore recommended. Conventional adiabatic MRS sequences (i.e. localization by adiabatic selective refocusing, LASER) have relatively long TEs, especially when optimized to measure the strongly coupled spins of citrate in the prostate. The semi-LASER (sLASER) sequence has a significantly shorter TE, although it does not provide adiabatic excitation. Therefore, we propose an adiabatic sLASER sequence that either has a composite adiabatic slice-selective excitation (cLASER) or a non-slice-selective adiabatic excitation (nsLASER), allowing for shorter TEs, whilst maintaining the adiabatic spin excitation. Furthermore, the spatial properties of the composite adiabatic excitation allow for a high slice excitation bandwidth, resulting in negligible chemical shift displacement artifacts. Exclusion of the slice selection can be considered once the field of view extends beyond the transmit field of the RF coil. The use of a transceiver at high magnetic field strengths has shown that the cLASER and nsLASER sequences are suitable for MRSI of the prostate in both phantom and in vivo validations.