Introduction

Parametric mapping provides valuable insight into myocardial tissue microstructure and T₁ and T₂ mapping abnormalities have been associated with fibrosis and edema (1). T_1ρ, the spin-lattice relaxation time in the rotating frame, is a complimentary technique that is sensitive to macromolecular content and thus may be more sensitive to collagen accumulation in myocardial infarctions (2). This may enable the characterization of myocardial scar without the need for gadolinium-based contrast agents. However, comprehensive characterization with all 3 techniques is often time consuming as they are generally performed in separate acquisitions. We propose an extension to the multi-parametric SAturation-recovery single-SHot Acquisition (mSASHA) technique (3) that provides co-registered T₁, T₂, and T_1ρ maps in either a single 13-heartbeat breath-hold or 51 heartbeats of free breathing.

Methods

The multi-parametric SASHA sequence consists of a series of single-shot images with various magnetization preparations (Fig. 1). The first image is unweighted, followed by images with saturation recovery (SR) T₁-weighting, SR plus T₂-preparation (SR+T₂p) weighting, and SR plus T_1ρ-preparation (SR+T_1ρ) weighting. T₂ weighting is achieved using an adiabatic T₂ preparation module with a specified echo time (TE). T_1ρ weighting is achieved using rectangular tip down/up pulses with 2 composite refocusing pulses and continuous wave excitation in between them with a specified spin-lock time (TSL).

The mSASHA sequence can be acquired in a single 13-heartbeat breath-hold (Fig. 1a) or as a free-breathing acquisition (Fig. 1b). With free breathing, the reference image is SR prepared with a long TS (>4 HBs) and the basic scheme is repeated 3 times, improving both imaging precision and patient comfort. High-contrast images were used improve the quality of motion correction (4).

T₁, T₂, and T_1ρ are simultaneously calculated from all images using a 4-parameter model (Fig. 2). The effect of linear k-space ordering on the signal intensity is modeled as a period of T₁ recovery, enabled by the use of the variable flip angle (VFA) modulation previously described (3,5).

The mSASHA prototype sequence was evaluated in a NiCl₂-doped agarose T1MES phantom (6) on a 1.5T system (MAGNETOM Sola, Siemens Healthcare, Erlangen, Germany) with comparison to gold standard spin-echo and commonly used SASHA, Modified Look-Locker Inversion recovery (MOLLI) and T₂p balanced steady-state free precession (T₂p-bSSFP) sequences. mSASHA was acquired with both breath-hold and free breathing protocols with a basic sampling scheme consisting of a reference image, 5 SR images (TS = 600 ms), 2 SR+T₂ images (TS = 600+RR ms, TE = 35, 55 ms), and 2 SR+T_1ρ images (TS = 600+RR ms, spin-lock time (TSL) = 35, 55 ms). Other sequence parameters were 360×270 mm² field of view, 1.4×1.9 mm² in-plane resolution, 8 mm slice thickness, 1.26/2.93 ms bSSFP TE/TR and variable flip angle with 100° maximum. SASHA, 5(3)3 MOLLI, and T₂p-bSSFP were acquired with flip angles of 70° VFA, 35° CFA, and 70° CFA respectively and other parameters matched.

The mSASHA, SASHA, MOLLI, and T₂p-bSSFP sequences were also evaluated in 3 volunteers (40±1 yrs, 1 female) in a single short-axis slice on both Sola and Aera 1.5T systems. The left ventricular myocardium was segmented avoiding regions with visible artifact and T₁, T₂, and T_1ρ values are reported as mean±standard deviation of pixels within the ROIs.

Results

In the NiCl₂-agarose T1MES phantom, breath-hold mSASHA T₁ and T₂ maps had good agreement with gold standard spin echo (0.3±0.4% and 2.7±1.9% error respectively) while reference MOLLI and T₂p-bSSFP had substantial errors (Fig. 3). mSASHA T_1ρ values were 6.1±0.8% higher than T₂. Free-breathing mSASHA T₁, T₂, and T_1ρ values were not statistically significantly different than breath-hold mSASHA values (p>0.05). Breath-hold mSASHA T₁ and T₂ maps had similar coefficient of variation (CoV) compared to reference MOLLI and T₂p-bSSFP respectively (p>0.05). Free breathing mSASHA reduced the CoV by 24±16%, 30±10%, and 36±12% in T₁, T₂, and T_1ρ respectively.

In-vivo breath-hold mSASHA T₁, T₂, and T_1ρ maps had good image quality with the exception of artifacts present in the inferolateral wall for T_1ρ, corresponding to a region of significant off-resonance (Fig. 4). Average myocardial mSASHA values of 1203±17 ms for T₁, 41.4±1.0 ms for T₂, and 46.9±0.3 ms for T_1ρ (Fig. 5). Free-breathing and breath-hold mSASHA values were not different for all parameters (p>0.05) and similar to T₁-only SASHA (p>0.05). In-vivo breath-hold mSASHA T₁ and T₂ pixel map precision was similar to MOLLI and T₂p-bSSFP techniques (p>0.05) and free-breathing mSASHA had lower myocardial CoV in all cases.

Discussion

The proposed mSASHA sequence provides simultaneous T₁, T₂, and T_1ρ maps in a single acquisition with either breath-hold or free-breathing protocols. Breath-hold mSASHA T₁ and T₂ maps had similar precision to conventional MOLLI and T₂p-bSSFP with superior accuracy in only a single breath-hold. mSASHA T_1ρ maps had good precision similar to T₂ maps, but off-resonance sensitivity of the T_1ρ preparation module requires additional optimization. The free-breathing mSASHA protocol improved precision and may also be preferable in patient populations where breath-holds are challenging.

Conclusion

mSASHA provides multi-parametric characterization of the myocardium in a single acquisition, potentially replacing 3 separate breath-holds for T₁, T₂, and T_1ρ mapping. Further study in patients with edema, focal fibrosis, and diffuse fibrosis are required to determine the sensitivity of T₁, T₂, and T_1ρ to various pathologies.

Acknowledgements

No acknowledgement found.