Accelerated free-breathing 3D whole-heart T2 mapping at 0.55 T

Background: Quantitative myocardial T2 mapping aids in detection of edema and inflammation in cardiomyopathies^1,2. Traditionally acquired in 2D under breath-holds, with anisotropic resolution and limited coverage of the heart, recent advancements propose a free-breathing 3D T2 mapping with motion-corrected under sampled signal-matched (MUST-T2) sequence to enable whole heart acquisition with high isotropic spatial resolution at 1.5T³. However, it has not been tested in low field MRI which could reduce costs and improved patient comfort⁴. This study investigates the feasibility of MUST-T2 for whole-heart tissue characterization at 0.55T in ~ 5 min.

Methods: MUST-T2 is a research sequence that uses fat suppressed balanced SSFP readout with an undersampled 3D VD-CASPR trajectory³. 2D image navigators (iNAVs) enable nonrigid respiratory motion-correction and 100% respiratory scan efficiency⁵. A non-selective saturation pulse is applied after the ECG R-wave, with a constant saturation time (T_SAT) set to the maximal available value. MLEV4 T2-preparation pulse with variable TEs (TET2prep = [0, 30, 60] ms) is performed, acquiring sequentially three T2-weighted volumes (Fig.1).

MUST-T2 was evaluated on a standardized T1MES phantom⁶ and in-vivo experiments in five healthy subjects using a 0.55T scanner (MAGNETOM Free.Max; Siemens Healthineers, Forchheim, Germany) with a 12-channel coil. Acquisition parameters include⁷: FOV= 190 mm, resolution= 2 mm³, TR/TE=4.76ms/2.42ms, acceleration factor of 4, total scan time = 5 min. For phantom comparison a 2D spin echo (SE) sequence (8 TEs from 10-640 ms) was acquired, and breath-held 2D T₂p-bSSFP maps were acquired in-vivo in short axis for basal, mid and apical slices. The three volumes were reconstructed simultaneously using HD-PROST ⁸ with T2 mapping performed via dictionary matching. The dictionary was generated using Bloch simulations with KomaMRI⁹. T2 values ranged from 10 ms to 250 ms, with 1% increments. Finally, dot-product matching was performed between the dictionary and the volumes to estimate the maps.

Results: MUST-T2 phantom maps at 0.55T show good agreement with SE reference values (R=0.99) in the range of interest (< 100 ms) (fig.2a). Representative T2 maps from three healthy subjects using 3D MUST-T2 and 2D T2p-bSSFP showed similar visualization of the left ventricle and surrounding structures (fig.2b). Bull's eye plots showed T2 mean values of 48.7 ± 4.3 ms for MUST-T2 and 57.7 ± 3.2 ms for T2p-bSSFP, with coefficients of variation of 8.8% and 5.5%, respectively. An underestimation of T2 values by MUST-T2 (mean difference = -8.99 ms) was observed as previous reports³.

Conclusion: In this work we demonstrate the feasibility of 3D MUST- T2 mapping sequence using a 0.55T contemporary MR scanner. The proposed approach achieves T2 values comparable to reference values for phantoms and promising results in-vivo. Future work will include evaluation in a larger cohort of healthy subjects and patients with acute myocardial infarction and myocarditis.

Figure 1. Schematic overview of the free-breathing 3D motion-corrected under sampled signal matched (MUST) T2 technique for whole-heart myocardial T2 mapping at 0.55T. Three T2-prepared volumes are acquired sequentially with increasing TET2prep ([0,30,60] ms). A nonselective saturation pulse is applied immediately after the electrocardiogram (ECG) R-wave to avoid recovery heartbeats. A 2D image navigator is acquired to enable nonrigid respiratory motion correction of the heart. All T2 prepared volumes are reconstructed simultaneously with high-dimensionality undersampled patch-based reconstruction (HD-PROST). A dictionary is then simulated and matched to the measured signal to generate the whole-heart T2 maps. Abbreviations: AW, acquisition window; bSSFP, balanced SSFP; iNAV, image-based navigator; TD, trigger delay; TSAT, saturation time.
Figure 2. (a). Phantom results for 3D MUST-T2 at 0.55T. Plots compare the mean T2 values derived from the 9 vials for MUST-T2 and T2p-bSSFP with the ground-truth T2 values (measured by spin echo [SE] with 8 TEs from 10-640 ms). (b). T2 maps obtained using the free-breathing 3D MUST-T2 and the conventional breath-held 2D T2p-SSFP sequences at 0.55T for 3 healthy subjects. The 3D MUST-T2 slices were reformatted to short axis to match the 2D T2 map acquisitions. Good visualization of the myocardium and surrounding structures can be observed on the 3D MUST-T2 maps. Acquisition times are expressed as minutes: seconds. Abbreviations: AT, acquisition time.
Figure 3. The T2 accuracy of the 3D MUST-T2 sequence at 0.55T versus conventional 2D T2p-SSFP, as measured by the mean T2 value, are shown in the left ventricular segmentation. The T2 values measured with 3D MUST-T2 are in good agreement with the literature (T2 = 48.7 ± 4.3 ms), showing underestimation with respect to 2D T2p-SSFP (3).