Table 1.
Imaging parameters of the B-TFE sequence for short-axis ventricular cine.
Figure 1.
RF homogeneity data acquisition and image post-processing.
Short axis cardiac modulus images (a, d) and B1 maps (b, e) acquired with single-source (a, b) and dual-source (d, e) RF excitation. The ROIs outlining the heart were placed on the modulus images (a, d), and then the same ROIs were copied to B1 maps (b, e) to evaluate the B1 homogeneity. Improved B1 homogeneity and reduced areas of dark spots were seen on (e) compared with (b). The distribution of the mean percentage of the achieved FA calculated from all pixels inside the ROI of the B1 maps was shown on figure c (without) and f (with dual-source RF excitation). The distribution of FA on (f) was more concentrated and uniform than that on (c). In addition, the average percentage of achieved FA on (f) was higher than that on (c).
Figure 2.
Quantitative assessment of RF field homogeneity.
Box-type diagrams showed average percentage of achieved FA (a) and coefficient variance (CV) (b) as a measurement of RF homogeneity for dual-source and single-source. A higher average percentage of FA and a lower mean CV was obtained by using of dual-source mode. Both parameters improved with dual-source from all subjects.
Figure 3.
Image contrast at B-TFE cine imaging.
Bar graphs showed CRs at LV-to-septum contrast (a) and RV-to-septum contrast (b) imaging with and without dual-source RF shimming. For CRs, the images with use of dual-source RF shimming (M1, M2, M3) revealed greater LV-to-septum and RV-to-septum versus the corresponding images with single-source mode. The CR of M2 was the largest with the FA = 58°. *** P<0.001 vs. single-source mode.
Table 2.
Comparison of the LV(RV)-to-septum contrast within the four groups.
Figure 4.
Artifacts on the short-axis B-TFE images of four groups.
The cine images acquired in groups M0 (a), M1 (b) and M2 (c) were all suffered from dark banding artifacts (thin arrows) seriously. Compared with group M0 (a), group M1 (b) reduced the dark banding artifacts to some extent and was better on the image homogeneity. Within the four groups, the CR of group M2 (c) was largest when the FA was increased, however, the off-resonance artifacts of group M2 (c) was pronounced in the ventricular wall (thin arrows). The off-resonance artifacts (thin arrows) of group M3 (d) were reduced and shifted away from the heart with dual-source RF shimming and a shortest TR, furthermore, group M3 (d) achieved good CR of blood-to myocardium contrast.
Table 3.
Comparison of the off-resonance artifacts scores within four groups.