Optimized saturation pulse rrains for SASHA T₁ mapping at 3T

Background

SASHA and MOLLI T₁ mapping sequences can have errors in calculated T₁ values when their magnetization preparation pulses do not fully saturate/invert magnetization [1, 2]. The commonly used 90°-90°-90° saturation pulse train [3] has poor performance at 3T due to large B₁ field inhomogeneities. We propose that a new hard RF pulse train with numerically optimized flip angles [4] will offer superior performance and reduce errors in SASHA T₁ values due to incomplete saturation.

Methods

Flip angles for a 6-pulse train were optimized by minimizing the maximum residual longitudinal magnetization in Bloch equation simulations performed over ranges of values expected at 3T: 40-120% B₁ scaling, -240-240 Hz off-resonance, 200-2000 ms T₁, and 14 μT B₁ strength. Complete spoiling of transverse magnetization was assumed during spoilers. Optimization code is available at https://bitbucket.org/kelvinc/pulsetrainopt.

Saturation performance for the 90°-90°-90° and the 6-pulse train was measured in a phantom with saturation recovery GRE. B₀ and B₁ maps were calculated using multi-TE and multiple flip angle GRE respectively. A magnetic field gradient was used to produce a range of off-resonance and experiments were repeated with the prescribed pulse train flip angles scaled by 40-120% to emulate B₁ inhomogeneity.

SASHA and MOLLI T₁ mapping were performed using investigational prototype sequences on a 66 kg swine in a 3T system (MAGNETOM Skyra, Siemens AG, Germany). SASHA was acquired using both the 90°-90°-90° and proposed pulse train with a 45° imaging flip angle. MOLLI used an optimized inversion pulse (2) with a 20° flip angle. A B₁ map was acquired using a saturated double angle method with single-shot EPI readouts.

Results

The optimized 6-pulse train flip angles were 115-90-125-85-176-223° with a 33 ms duration. The 6-pulse train had excellent performance (Fig. 1), with an average and maximum absolute residual longitudinal magnetization over the optimization range of 0.27% and 0.87% respectively. Experimental data had excellent agreement with simulations.

Figure 1

Simulated and experimentally measured residual longitudinal magnetization for a commonly used 90°-90°-90° saturation pulse train and a proposed 6 pulse train. White boxes denote the 3T optimization range.

In the swine study, the B₁ varied from 30-95% across the left ventricle (LV) profile (Fig. 2). MOLLI and 90°-90°-90° SASHA T₁ maps show a >50% artifactual decrease in T₁ values with reduced B₁ values in the lateral wall. SASHA T₁ values using the 6-pulse train are more spatially homogeneous (1386±70 ms across the entire LV profile).

Figure 2

A B₁ map (bottom left) and T₁ maps (top row) using SASHA with a reference 90°-90°-90° saturation pulse, SASHA with a proposed 6 pulse train, and MOLLI in a swine. A profile of T₁ values along the left ventricular wall shows decreased T₁ values in the lateral wall coinciding with reduced B₁ values.

Conclusions

A saturation pulse train optimized for B₀, B₁, and T₁ ranges expected at 3T was shown to have residual longitudinal magnetization of <1%. In-vivo swine MOLLI and SASHA data with the commonly used 90°-90°-90° pulses had >50% T₁ variation due to B₁ inhomogeneity while 6-pulse train SASHA had a 5% coefficient of variation.

Funding

Canadian Institutes of Health Research, Alberta Innovates - Health Solutions, NIH/NHLBI Intramural Research Program.