Skip to main content
SCMR 29th Annual Scientific Sessions Main banner
Final Program


Program-At-A-Glance

Oral Abstracts Session

Virtual Recording

Session: Oral Abstracts Science Session 5

Multiphase systolic and diastolic cardiac diffusion tensor imaging using higher order motion compensation at 300 mT/m gradient strength

Saturday, February 7, 2026
12:00 PM - 12:10 PM  
Location: Americas I and Hall 3 (2nd Lower Level)
Background: Cardiac diffusion tensor imaging (cDTI) enables non-invasive characterisation of myocardial microstructure by probing water diffusion in the heart tissue (1).

Myocardial structure and sheetlet orientations vary dynamically through the cardiac cycle. Ex vivo studies have shown differences in helix angle distributions and sheetlet orientation between systole and diastole (2–4). Similar findings have been observed in vivo using STimulated Echo Acquisition Mode (STEAM)-based cDTI (5).

In contrast, up to 2nd order motion compensated (M2) spin-echo cDTI is typically acquired in systole, as data quality is poorer during diastole (6). Higher order motion compensation (up to M4 and M6) was first demonstrated in systole (7,8). Afzali et al. extended this to M3 using an ultra-strong gradient (300 mT/m) MR system (9). Recently, Wen et al. explored diastolic cDTI with M3 but found signal dropouts in myocardium (10).

In this study, we establish the feasibility of dual-phase cDTI in the human heart in vivo using higher order (up to M6) motion-compensated spin echo sequences.

Methods: Cardiac DTI was performed on a Connectom 3T scanner (Siemens, Gmax = 300 mT/m) in ten healthy volunteers (age 20.1 ± 1.6 years; 6 females). A prototype cDTI sequence with ZOOM-EPI was used (11) (FOV = 320×120 mm²; resolution = 2.3 mm; slice thickness = 8 mm, three short-axis slices, ECG-gating, free-breathing, acquisition in systolic and diastolic phases). Diffusion gradient waveforms were designed with the NOW toolbox with M2, M4 and M6 motion compensation (12). TE was minimised and ranged from 56 to 76 ms, and b-values were 100 and 450 s/mm².

Data were registered, and diffusion tensors fitted to compute mean diffusivity (MD), fractional anisotropy (FA), helix angle (HA), and secondary eigenvector angle (E2A), respectively. Manual segmentation excluded regions with artefacts and / or distortions for global analysis.

Results: Figure 1 shows representative systolic and diastolic images across motion-compensation orders. In diastole, M2 images exhibited pronounced myocardial signal dropouts, whereas M4 and M6 improved image quality, with M6 showing a small SNR penalty from the longer TE. In systole, M2 provided highest SNR, but M4/M6 reduced subtle motion artefacts.



Figure 2 illustrates cDTI maps in systole and diastole. The average (over myocardial voxels) MD, FA and absolute E2A were calculated for each subject and Table 1 shows the mean ± standard deviation across subjects. Diastolic MD and FA were 48% and 23% higher in systole compared to diastole, consistent with suboptimal motion compensation. This effect was mitigated using M4 and M6.

Conclusion: Dual-phase spin echo cDTI with higher-order motion compensation on an ultra-strong gradient system was successfully implemented. This enabled more robust assessment of changes in myocardial architecture between systole and diastole.

Acknowledgements

JES and IT are joint senior authors.

Mean ± standard deviation of mean diffusivity (MD), fractional anisotropy (FA) and absolute secondary eigenvector angle (E2A) for second, fourth and sixth order motion compensation (M2, M4, and M6) for both systolic and diastolic phases, inside a left ventricular mask, averaged over ten volunteers.












Systole


MD [×10-3 mm2/s]


FA


E2A [degree]


M2


1.53 ± 0.06


0.31 ± 0.02


31 ± 4


M4


1.60 ± 0.07


0.30 ± 0.02


30 ± 5


M6


1.62 ± 0.06


0.30 ± 0.02


31 ± 3


Diastole


MD [×10-3 mm2/s]


FA


E2A [degree]


M2


2.26 ± 0.67


0.38 ± 0.12


30 ± 7


M4


1.73 ± 0.19


0.30 ± 0.02


21 ± 5


M6


1.71 ± 0.12


0.31 ± 0.03


21 ± 5

Figure 1: Representative diffusion-weighted images acquired using up to second, fourth, and sixth order motion compensation (M2, M4, and M6) for b = 100, and 450 s/mm2 in systolic and diastolic phases. Regions with artefactual signal dropout are indicated by red arrows.
Figure 2: Representative mean diffusivity (MD), fractional anisotropy (FA), helix angle (HA), and secondary eigenvector angle (E2A) maps using second, fourth, and sixth order motion compensation (M2, M4, and M6) in systolic and diastolic phases.