Synopsis

Resolving the dynamics of cardiac myofiber orientations with cDTI could have significant value in the assessment of cardiac disease. Unfortunately, cDTI measurements are not reliable at all cardiac phases and would require extremely long acquisitions. Fusing cDTI with time-resolved displacement maps from cine DENSE could enable evaluation of the dynamics of myofiber orientations using cDTI obtained at a single cardiac phase. The objective of this study was to generate dynamic myocardial fiber maps and validate the approach with dual-phase cDTI measurements.

Background

Cardiac Diffusion Tensor Imaging (cDTI) directly measures myocardial microstructure in vivo. Capturing diffusion tensor dynamics throughout the cardiac cycle has shown value in understanding conditions such as HCM and DCM ^1,2. Unfortunately, cDTI is typically limited to measurements at one or two cardiac phases. While recent developments in motion compensated diffusion encoding waveforms have improved cDTI robustness, precise and accurate measurements are not possible at all cardiac phases ^3,4. Furthermore, cDTI scans currently take several minutes per cardiac phase, making time-resolved measurements impractical.

Therefore, the objectives of this study were: 1) to generate time-resolved maps of myofiber orientations by combining cDTI data acquired at a single cardiac phase and cine DENSE displacement data; and 2) to validate these time-resolved myofiber maps using information from dual-phase cDTI.

Methods

Imaging: A single mid-ventricular short-axis slice was acquired with cDTI and spiral cine DENSE in healthy volunteers (N=9, 2 females, 7 males, aged 25-42) at 3T (Siemens Prisma) after obtaining informed consent. Imaging parameters for cDTI: free-breathing (navigator triggered) acquired at mid-systole and end-diastole with 2x2x5mm, TE=65ms, TR~3000ms, b-value=0 and 350s/mm², N_avg=10, N_dir=12, scan time=10min; DENSE: balanced 4-point encoding, navigator gated, 2.5x2.5x8mm, TE/TR=1.04/15, K_e=0.06cycles/mm, N_avg=3, spiral interleaves=10, scan time=5min with navigator gating.

Post-processing (Fig. 1): Lagrangian displacements were extracted from cine DENSE data with custom software ⁵. cDTI myofiber orientations were co-registered with the DENSE data onto a finite element (FE) mesh of the short-axis slice. Linear displacement interpolation was used to compute the continuum deformation mapping that describes cardiac motion through the cardiac cycle from a chosen reference configuration, here mid-systole (MS_REF) or end-diastole (ED_REF). The deformation gradient tensor F calculated from φ was used to transform the cDTI myofiber orientations acquired at two times: MS_REF and ED_REF to their respective computationally deformed configurations at end diastole, ED_DEF and mid-systole MS_DEF.

This permitted comparing the myofiber maps from the acquired ED_REF and MS_REF data to the computed ED_DEF and MS_DEF data to validate the mapping. It was also possible to compare computed fiber maps between methods at any intervening “synthetic” cardiac phase. Mesh elements were directly shared between REF and DEF fiber maps, therefore direct angular comparison between fiber vectors was possible at all phases. Helix angles (HA) were calculated for each mesh element of the ED_DEF and MS_DEF datasets and were binned by percent wall depth from epicardium (0%) to endocardium (100%) in increments of 10%. Transmural HA distributions were compared between maps at the mid-systolic and end-diastolic time points.

Results

Figure 2 shows the time-resolved myofiber vector maps deformed from both MS_REF and ED_REF. There was a mean angular difference of 27.2±17.5° between vectors across all phases. Between all myofibers acquired at ED_REF and those computed at ED_DEF there was a 24.8±18.3° difference. Comparing between all elements in the datasets at each “synthetic” cardiac phase, there was an overall 26.5±18.5° difference between fiber vectors computed with the MS_REF and ED_REF data (Figure 3).

Figure 4 shows transmural HAs across all volunteers, and representative myofiber vector maps colored by HA at the MS_REF or ED_REF acquisition time points. HA maps were largely in agreement between the two methods. Across all elements in the the mid-systolic (MS_REF vs. MS_DEF) or end-diastolic (ED_REF vs. ED_DEF) slices, HA’s differed by (Δθ=9.1±26.3° and 7.7±25.1°), respectively.

Discussion

Combining cDTI and cardiac DENSE data allows for time-resolved maps of myofiber orientation that cannot currently be easily acquired with cDTI alone. Myofiber maps derived independently from systolic and diastolic cDTI data show moderate agreement between fiber angles and better agreement between helix angle maps. Fiber vector agreement was within the cone of uncertainty ⁶ (data not shown), which measures the precision of cDTI fiber angle measurements, for diastolic phase cDTI measurements.

The difference between time-resolved fiber vectors computed with MS_REF and ED_REF datasets is greatest during systole. This may arise from the lack of LV twist in the displacement mapping model, which is greatest in systolic time points.

While HA increased transmurally, these values were lower than expected at the epicardial and endocardial border. This could be explained by the interpolation scheme used to register the cDTI fibers to the finite element mesh and conservative masking.

The comparison between MS_REF and ED_REF mappings validates our model by demonstrating that fibers from one acquired cDTI phase can be deformed through the cardiac cycle using DENSE to another time point within the precision expected for diastolic cDTI itself.

Conclusion

Acknowledgements

References

1. Ferreira PF, Kilner PJ, McGill L-A, et al. In vivo cardiovascular magnetic resonance diffusion tensor imaging shows evidence of abnormal myocardial laminar orientations and mobility in hypertrophic cardiomyopathy. Journal of Cardiovascular Magnetic Resonance 2014;16(1):1.

2. von Deuster C, Sammut E, Asner L, et al. Studying Dynamic Myofiber Aggregate Reorientation in Dilated Cardiomyopathy Using In Vivo Magnetic Resonance Diffusion Tensor Imaging. Circulation: Cardiovascular Imaging 2016;9(10):e005018.

3. Stoeck CT, Von Deuster C, Genet M, Atkinson D, Kozerke S. Second-order motion-compensated spin echo diffusion tensor imaging of the human heart. Magnetic resonance in medicine 2015.

4. Aliotta E, Wu HH, Ennis DB. Convex optimized diffusion encoding (CODE) gradient waveforms for minimum echo time and bulk motion–compensated diffusion-weighted MRI. Magnetic resonance in medicine 2016.

5. Spottiswoode BS, Zhong X, Hess A, et al. Tracking myocardial motion from cine DENSE images using spatiotemporal phase unwrapping and temporal fitting. IEEE Transactions on medical imaging 2007;26(1):15-30.

6. Jones DK. Determining and visualizing uncertainty in estimates of fiber orientation from diffusion tensor MRI. Magnetic Resonance in Medicine 2003;49(1):7-12.