1356505 - Multi-parametric T1, T2, and Adiabatic T1ρ Mapping with mSASHA in Healthy Volunteers at 1.5T and 3T

Quantitative myocardial mapping is commonly used to characterize tissue microstructure and native T₁ and T₂ have been associated with fibrosis and edema. T_1ρ has been proposed to be sensitive to collagen, enabling non-contrast detection of myocardial scar (1).  Recent adiabatic T_1ρ preparations have improved upon the previous sensitivity of conventional T_1ρ spin-lock pulses to B₀ and B₁ inhomogeneity (2).  The optimal T_1ρ pulse design parameters, performance, and normative values at 3T have yet to be established.

Multi-parametric sequences are ideal for comprehensive tissue characterization due to the reduction in acquisition time and intrinsic co-registration of maps.  Multi-parametric saturation-recovery single-shot acquisition (mSASHA) can be used for simultaneous T₁, T₂, and T_1ρ mapping in a single breath-hold (3, 4).  Conventional spin-lock and various adiabatic T_1ρ preparations are evaluated on healthy volunteers scanned at both 1.5T and 3T.

Methods:

The mSASHA sequence consists of a single non-prepared (anchor), 6 T₁-weighted, 2 T₁+T₂-weighted, and 2 T₁+T_1ρ-weighted images acquired in a 15-heartbeat breath-hold (Fig. 1).  High-contrast images were used to improve motion correction and T₁, T₂, and T_1ρ maps were calculated jointly from all images.  T_1ρ preparation was achieved using a pair of positive/negative frequency swept hyperbolic secant pulses with f_max=1600-1650 Hz and β=3.5, 5.0 and 7.0 (2).  Reference spin-lock T_1ρ preparation was also used with ω=400 Hz. All T_1ρ durations (TSL) were 60 ms.

Five healthy volunteers (41±1 yrs, 4 female) were imaged at both 1.5T and 3T (MAGNETOM Aera and Prisma, Siemens Healthcare, Erlangen, Germany) on the same day.  Mean and standard deviations over the left ventricular myocardium were calculated for each map.

Results:

mSASHA parametric maps visually had good image quality (Fig. 2), with T₁ and T₂ values in agreement with previously published T₁/T₂-only mSASHA (Table 1) (2).  T_1ρ with β=7.0 and conventional spin-lock exceeded energy limits in 1 subject each and these data were excluded.  Adiabatic T_1ρ values were higher than spin-lock T_1ρ with a strong dependence on β from 3.5 to 7.0 at both 1.5T and 3T (46% and 56% increase respectively).  No visible B₀ dependence of T_1ρ maps were observed at 1.5T, but spin-lock T_1ρ had significant artifacts in 2 subjects at 3T coincident with significant B₀ deviations (Fig. 2).  MOLLI T₁ values were also decreased in the off-resonance region while all mSASHA maps were unaffected.  Precision (Table 1) was improved for all parameters at 3T compared to 1.5T and similar to T₁/T₂-only mSASHA (3).

Conclusion:

mSASHA provides comprehensive T₁, T₂, and T_1ρ characterization in a single breath-hold with improved precision at 3T compared to 1.5T.  Adiabatic T_1ρ preparation improved robustness to B₀ inhomogeneity, enabling imaging at 3T.  Further clinical studies are required to determine the added value of T_1ρ in various cardiomyopathies and the optimal adiabatic pulse parameters for detection of myocardial scar.