SCMR 22nd Annual Scientific Sessions
Myocardial tissue health can be non-invasively assessed using quantitative techniques, such as T1 and T2 mapping. We propose an extension of the SAturation-recovery single-SHot Acquisition (SASHA) T1 mapping technique  to allow for efficient joint acquisition of multi-parametric T1 and T2 maps in free-breathing.
Free-breathing SASHA was adapted by adding 2 T1- and T2-weighted images after conventional SASHA's 7 T1-weighted images and a long saturation recovery time reference image (Fig. 1). This set of acquisitions was repeated 7 times during free-breathing. T1 and T2 values were calculated using least-squares fitting to a 3-parameter model.
Flip angle optimization was performed using a variable flip angle scheme (VFA)  with a maximum FA of 50º to 100º in a healthy volunteer to evaluate both signal-to-noise ratio (SNR) and blood-tissue contrast of corresponding maps. This was followed by sequence testing on 8 NiCl2 agarose phantoms with a range of T1 and T2 values, imaged using a MAGNETOM Prisma 3T (Siemens Healthcare, Erlangen, Germany). Prototype SASHA and T2-prepared mapping with linear balanced steady state free precession (T2p-bSSFP) and centric gradient recalled echo (T2p-GRE) were used as reference sequences. All images were reconstructed with 1.3-1.4 x 1.3-1.4mm2 in-plane image resolution. In-vivo feasibility testing was then performed in 11 clinical patients with various cardiomyopathies (6 male, 54±13yrs). Myocardial T1 and T2 values were measured using SASHA, T2p-bSSFP and joint T1/T2 sequences in a mid-ventricular short-axis slice. SASHA T1 and T2p-bSSFP sequences were acquired in 11- and 9-heartbeat breath-holds and the proposed joint T1/T2 sequence employed a 2-minute free-breathing acquisition.
Improved SNR and blood-tissue contrast in T1 and T2 maps was observed with larger FAs up to 100º - this value therefore used for phantom and in-vivo imaging. Phantom joint T1 and T2 values had good agreement with reference SASHA T1 and T2p-GRE sequences (-4±4 ms, -3±5 ms respectively, p>0.05) while T2p-bSSFP overestimated T2 values (15±10 ms, p=0.004) due to its linear readout (Fig. 2). Patient myocardial T1 values were 1563±42 ms and T2 values were 38±2 ms with the joint sequence. In-vivo joint T2 values were shorter than T2 p-bSSFP values (-3.9±1.1 ms, p = 0.007), consistent with phantom results. Joint T1 values were 2% longer than breath-hold SASHA T1 values (36.3±19.2 ms, p = 0.0002) in-vivo. The joint T1/T2 sequence demonstrated an 11% and 42% improvement in precision of T1 and T2 values compared to reference sequences (patient example, Fig. 3).
The proposed joint T1/T2 mapping sequence allows for complementary, co-registered T1 and T2 maps to be acquired during free-breathing with high accuracy and improved precision compared to conventional SASHA and T2p-bSSFP sequences.