Synopsis

MR fingerprinting based on a fast imaging with steady-state precession (MRF-FISP) is immune to banding artifacts due to the inhomogeneous field, and thus has been widely used for rapid quantification of tissue parameters. However, the MRF-FISP could still be affected by banding artifacts for high-field-strengths, insufficient shimming, and wide field-of-view scanners. In such cases, use of much stronger spoiler gradients could alleviate the banding artifacts, but it would increase the diffusion weighting of the spoiler gradients and lead to the underestimation of T2 values. In this study, we examined the effect of the diffusion weighting and showed the correction methods.

Introduction

Methods

The MRF acquisition was initiated with an adiabatic inversion pulse, which was followed by 300 successive FISP acquisition periods with varying flip angel (FA) and repetition time (TR). Echo time (TE) was held constant (5 ms). FA was varied pseudo-randomly and sinusoidally (5 to 90^o) and TR was varied using a Perlin noise pattern (20 to 30 ms). The slice thickness was 2.5 mm. The FOV was 25.6 cm × 25.6 cm. The Cartesian sampling³ was used for the MRF acquisition. The whole TR was 8 s.

To examine the effect of diffusion weighting, the moment of the spoiler gradient was varied. To correct the B₁ inhomogeneity, the B₁ dimension was added to MRF dictionaries, and the B₁ map measured with a double angle method was incorporated into the dictionary-matching process. To consider the diffusion effect, the ADC dimension was also added to the dictionaries. The MRF dictionaries with a range of T₁ (20-1500 ms), T₂ (10-1200 ms), B₁/B₁(90^o) (0.6-1.2), and ADC (0-2.6×10^-3mm²/s) were created using a home-written extended phase graph simulator. The dictionary was matched to the acquired MRF signal evolution profile using vector-based inner product comparisons.

To correct the diffusion weighting effect, we tested two strategies. First, the ADC map was measured with a standard Stejskal-Tanner DW-SE sequence and incorporated into the matching process. Second, the echo signals were also recorded in the MRF-FISP acquisition. This method is referred to as the MRF based on a double-echo steady state (MRF-DESS)⁴. For MRF-DESS, to increase the diffusion sensitivity, monopolar diffusion gradients with varying gradient moments (16, 32, 48, 64, and 80π/2.5 mm) were inserted between the FID and echo.

The above methods were tested for nine test tubes filled with water with CuSO₄ to alter T₁ and T₂ and with glycerol to alter the diffusion constant. All measurements were performed on a vertical wide bore 4.74 T system (Oxford Instruments, bore diameter = 88.3 mm) with shielded gradients (MRTechnology, Japan).

Results and discussion

Figure 1(a) shows T₁ and T₂ maps of MRF-FISP without considering the diffusion. The MRF-T₁ estimates were independent of the spoiler gradient moment. For the smallest moment (4π/2.5 mm), the banding artifacts appeared in the T₂ map, as in the bSSFP image (Fig. 1(d)). The banding artifacts gradually disappeared as the moment increased, but the MRF-T₂ values decreased and deviated (Fig. 1(f)) from their standards (Fig. 1(b)).

This T₂ underestimation was corrected by incorporating the ADC map into the dictionary-matching process (Fig. 2(a)). The banding artifacts still appeared for the smaller moments, but they disappeared for the larger moments. The MRF-T₁ and T₂ estimates were almost the same as those measured with the standard method (Figs. 2(b) and 2(c)).

As such, the diffusion-weighting effect in MRF-FISP can be alleviated largely, but the correction requires the additional ADC experiment. This is not the case for MRF-DESS, which is the second strategy for avoiding the diffusion weighting. The T₁, T₂, and ADC maps generated from the MRF-DESS (Fig. 3) were in agreement with their standards. The MRF-DESS allows the simultaneous quantitative mapping of the ADC as well as T₁, T₂, and proton density, and thus is more efficient.

Overall, the both correction methods compensated for the T₂ underestimation largely, but there were slight deviations in the T₂ values from the standards. This would be caused by the insufficient sensitivities of the current sequences to T₂ and ADC. This limitation could be overcome by the suitable optimization of the sequences.

Conclusion

Acknowledgements

References

1. Ma D, et al. Magnetic resonance fingerprinting, Nature 2013;495:187-193.

2. Jiang Y, et al. MR Fingerprinting Using Fast Imaging with Steady State Precession (FISP) with Spiral Readout, Magn. Reson. Med. 2015;74(6):1621-31.

3. Gao Y, et al. Preclinical MR fingerprinting (MRF) at 7 T: effective quantitative imaging for rodent disease models, NMR Biomed. 2015;28:384-394.

4. Jiang Y, et al. Simultaneous T1, T2, Diffusion and Proton Density Quantification with MR Fingerprinting, Proc. Intl. Soc. Mag. Reson. Med. 2014;22:0028.