Spatially resolved quantification of the longitudinal relaxation time T₁ has recently emerged as an important quantitative biomarker with promising potential for non-invasive tissue characterization, especially in the presence of diffuse myocardial fibrosis [1]. The combination with post-contrast T₁-time and hematocrit enables the estimation of the extracellular volume fraction (ECV), which was shown to indicate fibrotic remodeling [2].

However, partial-volume effects near the highly intense blood pool signal corrupt the methods’ quantification accuracies and impair their reproducibility. Thus, data acquisition during systole has been proposed for inversion-recovery (IR) techniques [3,4], in order to benefit from the increased myocardial wall thickness. Saturation-recovery (SR) T₁-mapping is known to provide more accurate T₁-values and better resilience towards heart-rate variability and sequence variations at 1.5T [5].

Hence, in this work we sought to study the feasibility of systolic myocardial T₁- and ECV-mapping with a SR technique at 3T.

10 healthy volunteers (5m, 5f; 25±4y) underwent T₁-mapping before and 15 min after injection of a Gd based contrast agent (0.2 mmol/kg Dotarem; Guerbet, Aulnay-sous-Bois, France) at a 3T MRI scanner (Magnetom Skyra; Siemens Healthcare, Erlangen, Germany) with a 30 channel receiver coil array.

The T₁-mapping sequence comprised a Saturation-Pulse Prepared Heart-rate independent Inversion Recovery (SAPPHIRE) magnetization preparation [5], adapted for systolic acquisition as explained in Fig.1. Systolic T₁-maps were compared to conventional SAPPHIRE T₁-maps acquired during diastole. A WET module [6] was used for saturation and an adiabatic full passage tan/tanh pulse [7] for inversion, followed by a single-shot ECG-triggered bSSFP readout: TR/TE/α=2.6ms/1.0ms/35°, in-plane resolution=1.7×1.7mm², slice-thickness=6mm, field-of-view=440×375mm², bandwidth=1085Hz/px, #k-space lines=139, linear profile ordering, startup-pulses=5 Kaiser-Bessel, GRAPPA-factor=2.

T₁-weighted images were motion corrected with MoCo (Advanced Retrospective Technique; Siemens Healthcare, Erlangen, Germany). The thickness of the myocardium was evaluated as the area between the LV endo- end epicardial borders. T₁- and ECV values between systole and diastole, were statistically compared using a paired student's t-test at a significance level of p<0.05.

Fig. 2 depicts representative systolic T₁-maps. As illustrated, good visual quality, with a homogenous myocardium and sharp delineation towards the blood pool were obtained throughout the study.

Fig. 3 visualizes the effects of partial-voluming in a cross-section plot in the septal region of the left ventricle. In the diastolic T₁-maps, major T₁-time elevation at the myocardial blood interface can be observed, leaving only a small plateau in the center of the myocardium unaffected of partial-voluming. The increased myocardial thickness in the systolic images shows a larger plateau for evaluation of myocardial T₁-times. An increase in apparent myocardial thickness in the T₁-maps during systole (apical: 255±85%; mid-ventricular: 254±63%; basal: 209±40%) compared to diastole has been measured in average over all volunteers.

Fig. 4 shows the average pre-contrast T₁-values, T₁-time precision and ECV values of all 10 volunteers in AHA-16-segment bullseye plots. Systolic SAPPHIRE T₁-times (1554±83 ms) are significantly lower than diastolic T₁-times (1581±36 ms) (p<10^-7). Systolic ECV values (0.29±0.03) are significantly lower than diastolic ECV values (0.31±0.04) (p<10^-7).

Systolic T₁-mapping and ECV-mapping at 3T is feasible with the saturation recovery T₁-mapping technique SAPPHIRE and provides robust image quality. Shorter T₁-times during systole are most likely to be explained by a reduction of partial-voluming, achieved by an increase in myocardial thickness, as indicated by the comparable plateau values in Fig 3. Difference between systolic T₁- and ECV-values is significant despite the inter-subject variability. This finding is in accordance with a ShMOLLI (Shortened MOLLI) study comparing diastolic and systolic T₁-values [4]. The reported T₁-times are in good agreement with a recent study on saturation recovery at 3T (SASHA: 1431-1491 ms) [8].

Saturation-recovery T_1-mapping is inherently robust to changes in the R-R interval. Therefore, the proposed sequence design could potentially be used for HR insensitive T_1-mapping in arrhythmogenic patients, with no susceptibility to imaging artifacts induced by the major variations in the duration of the diastolic rest-period. Future studies in this cohort are warranted. In conclusion, our results show that SR T₁-mapping in systole might be an alternative to derive T₁- and ECV-values with reduced effects of partial volume.