Chemical exchange saturation transfer (CEST) and the closely related off-resonance T_1ρ methods are gaining popularity. However, quantitative CEST/T_1ρ imaging can be time consuming since the experimental repetition with different saturation lengths/powers is needed. Recently, a new way to acquire CEST/T_1ρ data was proposed¹ based on the exchange-sensitive properties of balanced steady state free precession sequence (bSSFPX).

The bSSFP (or TrueFISP) is a versatile sequence^2-4. The bSSFP sequence can be viewed as generating an effective off-resonance field analogous to CW and possessing similar spin-lock and saturation properties^1,5. In the previous study, the effective field approach was derived and proof-of-concept was demonstrated in steady-state conditions. Here, we implement the transient bSSFPX. The sequence has features similar to Look-Locker⁶ and IR-TrueFISP⁷. However, the current sequence does not require any additional preparation pulses and specifically focuses on saturation and off-resonance effects. The motivation is to acquire the transient information in a single shot, thus paving the way to fast, quantitative CEST/T_1ρ imaging.

All of the experiments were performed on a 3T MRI scanner (Ingenia, Philips Healthcare) using a 15-channel head-spine coil at room temperature. The pulse program was modified to acquire images continuously during the preparation echoes in a standard bSSFP sequence as shown in Fig.1. Thus the sequence acquires all, transient and steady-state images. Note that the bSSPFX acquires an XY-spectrum and not a Z-spectrum. The transient bSSFPX sequence was tested in two phantoms (i) 2% agar with 0.14 mM gadolinium and (ii) 10, 50 and 100 mM choline water solutions. The imaging parameters were TR/TE=2.0/1.0 ms, voxel size=3*3*8 mm, FA=40^o, half alpha startup echoes. 51 points in the XY-spectrum¹ were acquired at the frequency range ±500 Hz. 20 images (corresponding to different saturation times) were acquired at each off-resonance frequency. Since the number of phase encoding lines was 73 and the linear ordering was used, the effective saturation time for the i^th image t_sat=[37+(i-1)*73]*TR. Images were processed on a pixel-by-pixel basis using custom Matlab routines. The B₀ inhomogeneities and drifts were corrected using a procedure similar to the one described in Ref.11. MTR_asym was calculated and normalized to the upfield side of the XY-spectrum.

The evolution of the transient signal is governed by the relaxation times T_1ρ and T_2ρ, parallel and perpendicular to the effective field, respectively⁸ and by the precession frequency around the effective field¹. Similar to the CW case, both relaxation times depend on the off-resonance value as well as on the flip angle^1,8,9 and contain exchange contributions.

Fig.2a shows the evolution of agar XY-spectra from the transient state (lighter) into the steady state (darker) as the effective saturation time increases. In Fig.2b, the signals at +1.2 and -1.2 ppm are plotted against t_sat. The signals decay exponentially with T_1ρ as they approach the steady-state. Agar has no exchanging moieties, hence its steady-state bSSFP profile is symmetric around water and the steady-state signals are similar for both frequencies. Moreover, the measured T_1ρ at +1.2 and -1.2 ppm are similar (360±30 and 420±20 ms, respectively). Since the intrinsic T₂ in agar is short, the signal is largely governed by the T_1ρ decay and no transient oscillations were observed.

Fig.3 shows the 10 and 100 mM choline XY-spectra against effective saturation time. The T₁ and T₂ of the aqueous solutions are very long and prominent transient oscillations and spurious effects are observed, especially at the earlier time points. However, as the saturation time increases the profiles approach steady-state values. The steady-state profiles are asymmetric around water due to chemical exchange at ~1 ppm downfield (Fig.3b, arrow). The MTR_asym (1 ppm) also approaches its steady state values of ~2%, 10% and 16% for 10, 50 and 100 mM respectively (Fig.4). After initial transient oscillations (asterisks in Fig.4) the MTR_asym ­demonstrates exponential growth similar to QUEST curves¹⁰. Some residual oscillations due to transient effects or experimental imperfections are still observed. As shown in Fig.1, the acquisition is effectively performed during the saturation; hence one “shot” of the transient bSSFPX sequence acquires sufficient data for T_1ρ and/or exchange rate quantification, which may largely speed up the quantification process. The T_1ρ and, potentially, the exchange rate could be then obtained by fitting the curves (Fig.2b or Fig.4).

We have presented experimental implementation of transient bSSFPX measurements, aimed at the single-shot quantitative CEST/T_1ρ measurements. The sequence has a potential to quickly acquire quantitative information. Work is in progress setting up an analytical model for the signal evolution and extracting quantitative information (i.e. exchange rate) from the profiles.