Saturation-Recovery Myocardial T1-Mapping during Systole: Accurate and Robust Quantification in the Presence of Arrhythmia

Figure 1. Sequence diagrams of the systolic (a) and diastolic (b) SAPPHIRE T1-mapping sequence with ten readouts, as used in phantom and in healthy volunteers. Systolic T1-mapping starts with one image (IMG) without preparation, the following four images (seven images in patients) are preceded by a saturation pulse (SAT) in the heart beat before imaging, directly after the previous image acquisition. The remaining images (four in healthy, seven in patients) are acquired with an additional inversion-pulse (INV) with variable delay after the R-wave. (b) In diastolic (conventional) SAPPHIRE, the first image acquisition is also performed without magnetization preparation. However, for the remaining images, both the saturation- and the inversion-pulse are played within the same heartbeat before image acquisition. The image acquisition window is longer compared with systole.

Figure 2. Influence of arrhythmia studied in (a) a numerical simulation and (b) in phantom measurements. (a) Simulated influence of arrhythmia on diastolic T1-map quality is shown on a model mid-ventricular short-axis view and the corresponding standard deviation map. Blurring at the endo- and epicardial borders increased with increasing number of mis-triggered base-images. (b) Influence of simulated heart-rate variations on SAPPHIRE T1-values for seven phantom vials covering a broad T1-range. Circles indicate the T1-results with systolic SAPPHIRE for four different arrhythmia factors (defined as the standard deviation from the Gaussian distribution of a mean RR-length of 1125 ms). Reference diastolic T1-values are given as solid lines. No major influence of arrhythmia on systolic T1-values was found.

Figure 3. Reduction of partial-volume effects by higher myocardial thickness in systole. (a) Cross-sections through the LV septum in apical, mid-ventricular and basal short axis T1-maps, acquired with systolic and diastolic SAPPHIRE. The cross-sections through diastolic T1-maps (solid lines) show strongly elevated T1-times at endo- and epicardial borders, whereas in systole (dotted), this effect is reduced. Corresponding T1-maps of a healthy volunteer (m, 23 y) are shown in (b). The apparent myocardial thickness is clearly higher in systole compared to diastole, so more myocardial voxels can be included into T1-estimation without risking an elevation of T1-times by partial-volume effects from the highly intense blood-pool. (c) Correlation of systolic and diastolic native T1 in ten healthy subjects, each point indicating the average over the three slices in each subject, revealing a strong positive correlation of the two methods. The identity line is indicated in dashed blue, the solid green line represents best fit.

Figure 4. Indicative substudy for the comparison of MOLLI and SAPPHIRE sequence variants. Mean values of five healthy volunteers (3 f, 26 ± 3 y) in three short-axis slices (A = apical, M = mid-ventricular, B = basal) are displayed as bullseye plots (AHA-16-segment-model) for native myocardial T1-times and T1-time precision. They have been acquired with different sequence variants of the MOLLI and the SAPPHIRE T1-mapping method. The average across all segments is given in the bullseye centers, slice averages in the boxes below.

Figure 5. The left side shows native and post-contrast T1-maps of a healthy volunteer (m, 35 y), acquired with systolic (a) and diastolic (b) T1-mapping. T1-map quality is visually high in short axis apical (left), mid-ventricular (middle) and basal (right) slices. Mean values of ten healthy volunteers (5 m, 25 ± 4 y) in three short-axis slices (A = apical, M = mid-ventricular, B = basal) are displayed as bullseye plots (AHA-16-segment-model) for native myocardial T1- times, T1-time precision and ECV, acquired with the systolic SAPPHIRE technique (a) and the diastolic SAPPHIRE technique (b). The average across all segments is given in the bullseye centers, slice averages in the boxes below.

Figure 6. SAPPHIRE T1-mapping data in systolic (a) and diastolic (b) acquisition of a patient (m, 73 y) suffering from arrhythmia. Left ventricular myocardial borders are delineated in red in all 15 recovery images for the myocardial borders of the first cardiac frame, in yellow for all the following frames. In systole, the borders accord with the first frame, which reflects in less artifacts on the T1-map on the right. The detected 15 RR-lengths during that measurement are depicted below (c) to visualize the arrhythmia (variability in RR length (RR mean ± std) = 816 ± 134 ms; min: 613 ms; max: 1068 ms). (d) Boxplots showing native T1-estimations with systolic and diastolic SAPPHIRE of six patients (4 m, 52 ± 19 y) suffering from arrhythmia and ten healthy volunteers (5 m, 25 ± 4 y). Superimposed scattered data points indicate mean T1 per subject.